Cygnar v. City of Chicago

COFFEY, Circuit Judge. The plaintiffs-appellants, 13 Chicago police officers, appeal the district court’s grant of judgment notwithstanding the ver-diet (JNOV) in favor of the defendants-ap-pellees, City of Chicago, Fred Rice, and Raleigh Mathis. The court’s JNOV order overturned the lion’s share of a jury’s $4.29 million verdict finding that the plaintiffs had been transferred from the City's Office of Municipal Investigations to assignments within the Chicago police department because of their race and political affiliation. The plaintiffs further appeal the court’s entry of a directed verdict in favor of the defendant Harold Washington (the former mayor of Chicago). We affirm in part, reverse in part, and remand for further proceedings. FACTS Prior to the summer of 1984, the plaintiffs, thirteen white Chicago police officers, had all been assigned to posts within the City of Chicago's Office of Municipal Investigation (OMI). The mayor of Chicago established OMI, an independent division of city government, in 1977 (under the title “Office of Professional Review”) for the purpose of investigating corruption within city government. In 1980, Mayor Jane Byrne appointed James Maurer (defendant Mathis’s predecessor) to head the OMI unit. Maurer, with the benefit of an increased budget, selected and trained highly qualified officers of his choosing in an attempt to professionalize the office. Among these officers were all of the 13 plaintiffs, most of whom Maurer had known and worked with for many years.1 By the time of Maurer’s removal in April of 1984, OMI employed 32 sworn (Chicago police department) and unsworn (civilian) investigators. — 831 — During the 1983 mayoral campaign, which pitted Harold Washington against Jane Byrne in the Democratic primary and later Washington against Bernard Epton, the Republican candidate, in the general election, eight of the appellants campaigned for either Byrne or Epton. Several attended fund raisers for one or the other. One plaintiff, Shanahan, organized and co-hosted a large fund raiser for Byrne, which 1500 people attended. Five of the plaintiffs made campaign contributions to Epton, and another (Capesius) purchased a ticket and attended an Epton rally. Another plaintiff (Flanagan) canvassed his precinct and worked actively for Epton (whether Mathis knew of these activities was, as we shall see, one bone of contention at trial). The year following his election as mayor of the City, Washington removed Maurer, and on April 23, 1984, appointed Mathis to the position of Executive Director of OMI. Mathis, a strong Washington supporter, reported directly to William Ware, then Washington’s chief of staff. Shortly after his appointment, Mathis conducted a survey of the gender and race of the OMI staff, and on May 24, 1984, wrote a memo to Ware reporting that of the 32 police officers assigned to OMI, 28 were white males, three were black, and one, who had just started a month earlier, was a Hispanic female. Mathis’s memo also noted that all of OMI’s lieutenants and sergeants were white males. The memo concluded: “As this data indicates, affirmative action concepts do not exist in this department, either in its civilian or police components. Due to the obvious constraints in dealing with the civilian career service personnel, resolving this imbalance will be a long-range goal. However, in reference to the police personnel, this is being dealt with immediately. Further reports on this progress will follow.” Subsequently, Mathis met with Ware at City Hall to discuss the May 24 memorandum. During their discussion, Ware asked Mathis what he intended to do about the perceived “imbalance” in the racial makeup of OMI personnel. Mathis replied that “I told him I intend to correct that imbalance by the selection of different personnel from the police department.” Within weeks of his meeting with Ware, Mathis began instituting the transfer of sworn personnel from OMI to other positions within the Chicago police department. In mid-June 1984, Mathis informed three supervisors, including plaintiffs Di Maggio and Cappitelli, that he was transferring them out of OMI. Officer Di Maggio testified that Mathis told them he was transferring them “because I’m bringing my own people in.” Shortly thereafter, Paul Lewis, then OMI director of operations, informed plaintiff Capesius, another supervisor, that “I had better put my P.A.R. [Personnel Action Request] form in and get out or I could be dumped anywhere in the city.” Subsequently, Capesius requested and was granted a transfer. Similarly, plaintiff Wojnar testified that his supervisors advised him that the detectives’ positions were not stable and that his “status as a detective could not be guaranteed.” Based upon this information, Wojnar asked to be transferred to his previous unit in order to avoid “being sent to a distant location.” Plaintiff Gartner also requested a transfer in June but only after Mathis informed him that his transfer was imminent. About one month later (on July 25, 1984), Mathis ordered the transfer of a group of eight sworn investigators out of OMI. Five of the plaintiffs (Cygnar, Flanagan, O’Driscoll, Shanahan and Rubino) were transferred as part of this group. The other three plaintiffs (Bleke, Riggio and Murray) left OMI shortly thereafter at their own request although each testified that the July 25 transfer order played a role in his decision to request a transfer.2 — 832 — As the plaintiffs transferred out and Mathis brought in new officers to replace them, the racial and gender makeup of OMI personnel dramatically altered. Most new officers were not individual replacements for the plaintiffs although Flanagan testified that he was replaced by a black male, and Murray testified that he was replaced by a white male. While a number of the replacements were white, a majority of the people Mathis brought in were black or Hispanic. By August 31,1984, the number of white male officers at OMI had been reduced from 28 to 15 (Mathis brought in 13 of these 15) while the number of minority investigators had been increased from four to 17. Mathis reported these personnel changes, including the race and sex figures, to Ware in a memo dated August 30, 1984. Subsequently, the plaintiffs sued Mathis, Mayor Washington, Fred Rice, and the City of Chicago under 42 U.S.C. § 1983, alleging that their transfers out of the OMI unit had been politically and racially motivated in violation of the first and fourteenth amendments to the U.S. Constitution, respectively. Plaintiffs’ complaint also stated claims for relief under the due process clause of the fourteenth amendment and Title YII of the Civil Rights Act of 1964. A problem that plagues this case (to the defendants’ detriment) took shape soon after the plaintiffs filed their lawsuit — a problem that relates to the defendants’ defensive posture on the plaintiffs’ racial discrimination claim. Beginning with their answer and continuing throughout trial, the defendants have defended against the plaintiffs’ race claim on one primary ground: that Mathis’s actions were based not upon discriminatory animus but rather his legitimate desire to “bring in his own people.” But also thrown in for good measure have been hints at a backup theory: that if Mathis’s transfer decisions were based upon the plaintiffs’ race, Mathis still did not violate the plaintiffs’ fourteenth amendment rights because he was attempting in good faith to institute a valid affirmative action program. As we discuss later, this backup theory (essentially legal discrimination) is at odds with the defendants’ primary defense of non-discrimination. More significantly, the defendants failed to pursue this second line of attack in those terms either before or during trial — they neither filed a motion of summary judgment on the issue (couched in terms of qualified immunity or otherwise) nor did they raise the issue in their motion for a directed verdict at the close of the plaintiffs’ case.3 While the affirmative action defense should thus have been dead and buried at that point, the trial judge, at neither party’s request, resurrected the issue at the last minute in the form of this special interrogatory to the jury (which all the jurors answered in the affirmative).4 “Do you find that Raleigh Mathis’s decision to reassign plaintiff out of OMI to the extent it was substantially motivated by the intent to discriminate against plaintiff because of his race, was perceived by Mathis as a means to correct what he considered a previously-existing racial imbalance within OMI?” Taking the bait, the defendants finally included the lost “affirmative action de — 833 — fense” in their motions for a JNOV, which leads to the subject of this appeal: the district court’s ruling on the defendants’ motions for judgment notwithstanding the jury’s verdict in the plaintiffs’ favor. At the close of all the evidence, the jury (after deliberation) returned a verdict in favor of the plaintiffs on their race and political discrimination claims5 totalling $4.29 million. The jury’s verdict for each plaintiff was identical: $55,000 in compensatory and $275,000 in punitive damages (the latter against Mathis alone). As part of their verdict, the jury answered “Yes” to each of two special interrogatories: “1. Was Raleigh Mathis’s decision to reassign any of the following plaintiffs out of the Office of Municipal Investigations substantially motivated by the political affiliations of that plaintiff? 2. Was Raleigh Mathis’s decision to reassign any of the following plaintiffs out of the Office of Municipal Investigations substantially motivated by the race of that plaintiff?” As to each plaintiff, the jury answered “No” to a third special interrogatory: “3. If you have found either the political affiliation or race, or both, of any plaintiff was or were a substantial motivating factor or factors in Raleigh Mathis’s decision to reassign that plaintiff, would Raleigh Mathis have reached the same decision to reassign that plaintiff even in the absence of that plaintiff’s political affiliation and even without reference to that plaintiff’s race?” Finally, the jury answered “Yes” to the fourth interrogatory — again as to each plaintiff: “4. Do you find that Raleigh Mathis’s decision to reassign plaintiff out of OMI, to the extent it was substantially motivated by the intent to discriminate against plaintiff because of his race, was perceived by Mathis as a means to correct what he considered a previously-existing racial imbalance within OMI?” After the jury returned its verdict, the defendants moved alternatively for JNOV, a new trial, or remittitur. In these motions, the defendants argued that the evidence was insufficient to support the jury’s verdict, that Mathis was entitled to qualified immunity, and that both the compensatory and punitive damage awards were excessive and unsupported in the evidence. The trial court disposed of these motions in two separate orders which the court entered a month apart. In his first order, the trial judge (1) granted JNOV in favor of the defendants on the plaintiffs’ political discrimination claim, finding that “[n]o ... credence can be given to the jury’s determination as to Mathis’s having been politically motivated”; (2) granted JNOV against three of the plaintiffs (Bleke, Murray and Riggio) on their race discrimination claims but denied the motion as to plaintiffs Cap-pitelli and Di Maggio; (3) ruled that Mathis was entitled to qualified immunity; and (4) granted any successful plaintiff the option of remittitur (from $55,000 to $15,000) or a new trial on the issue of damages. Notwithstanding having already closely scrutinized the jury’s factual findings, the court withheld final ruling on the motions with regard to the remaining plaintiffs pending further factual submissions from the parties. The court stated that it was doing so for two primary reasons: (1) because it believed the parties’ previous submissions had failed to pay close enough attention to the legal significance of the factual differences among the plaintiffs; and (2) because the parties had failed to furnish the judge with the trial transcript. (In making his rulings, the judge stated that he “has had to rely on [his] own necessarily incomplete (though extensive) trial notes.” This statement is most surprising in that if the parties did not provide a transcript, we fail to — 834 — understand why the judge himself refused to order a transcript on his own before taking the drastic step of overturning a jury’s verdict in a complex case.) After receiving further submissions from the parties, and without the benefit of a trial transcript,6 the court entered its final order on the defendants’ post-trial motions. The order granted the defendants’ JNOV motion with respect to the remaining plaintiffs’ claims of race discrimination, except the plaintiff Flanagan, who was directly replaced by a black officer. As with plaintiffs Cappitelli and Di Maggio, the court granted Flanagan the option of accepting remittitur in the amount of $15,000 or a new trial on damages. Finally, the court ordered that “if and to the extent any of these rulings7 granting JNOV may be reversed on appeal, a new trial is awarded as to both liability and damages — unless, again at each plaintiff’s option, a like remit-titur to $15,000 is accepted.” The plaintiffs appeal the court’s entry of JNOV (to the extent the rulings are not in their favor) arguing, inter alia, that the court erred in substituting its judgment for that of the jury on the political and racial discrimination claims. Further, the plaintiffs insist that Mathis is not, as the court ruled, entitled to qualified immunity (and thus shielded from an award of punitive damages). The plaintiffs also take issue with the court’s conditional reduction of the jury’s compensatory damages award, as well as with the court’s entry of a directed verdict in favor of the defendant Washington. STANDARD OF REVIEW “The standard under which we review the district court’s decision to enter a directed verdict is the same as the standard used to evaluate a judgment notwithstanding the verdict, and is the same on appeal as it is in the trial court.” Eggert v. Weisz, 839 F.2d 1261, 1263 (7th Cir.1988). Reviewing the district court’s decision on the defendants’ motions for JNOV (and the dh rected verdict as to defendant Washington); we must determine de novo “whether the evidence presented, combined with all reasonable inferences permissibly drawn therefrom, is sufficient to support the verdict when viewed in a light most favorable to the party against whom the motion is directed.” Tice v. Lampert Yards, Inc., 761 F.2d 1210, 1213 (7th Cir.1985). See also Webb v. City of Chester, 813 F.2d 824, 827-28 (7th Cir.1987). In applying this standard, neither the district judge nor the appellate court may “resolve conflicts in testimony or weigh and evaluate the evidence, functions that are reserved to the factfinder.... If the evidence, taken as a whole, provides a sufficient probative basis upon which a jury could reasonably reach a verdict, without speculation over legally unfounded claims, the motion should be denied.” Anderson v. Gutschenritter, 836 F.2d 346, 348 (7th Cir.1988). In the words of the Supreme Court: “In every case, before the evidence is left to the jury, there is a preliminary question for the judge, not whether there is literally no evidence, but whether there is any upon which a jury could properly proceed to find a verdict for the party producing it, upon whom the onus of proof is imposed.” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 250-51, 106 S.Ct. 2505, 2511, 91 L.Ed.2d 202 (1986) (emphasis in original). — 835 — A different standard applies to our review of the trial court’s rulings on the defendants’ motions for a new trial (which the court granted in the event of reversal on the JNOY order). “The test to be applied in determining whether a motion for a new trial should be granted is whether ‘the verdict is against the weight of the evidence, that the damages are excessive, or that, for other reasons, the trial was not fair to the party moving.’ ” General Foam Fabricators v. Tenneco Chemicals, Inc., 695 F.2d 281, 288 (7th Cir.1982) (quoting Montgomery Ward & Co. v. Duncan, 311 U.S. 243, 251, 61 S.Ct. 189, 194, 85 L.Ed. 147 (1940)). Moreover, “[b]ecause the authority to grant a new trial .'.. is confided almost entirely to the exercise of discretion on the part of the trial court, ... the grant or denial of a motion for a new trial is not subject to review by this court, except upon exceptional circumstances showing a clear abuse of discretion.” General Foam, 695 F.2d at 288 (citations omitted). It is against this framework that we review the trial court’s rulings on the defendants’ post-trial motions. Initially, we turn to the question of racial discrimination. RACIAL DISCRIMINATION The district court properly concluded that based on the evidence, “the jury could rationally have decided, as it did, that Mathis’s decision to reassign plaintiffs out of the OMI unit (1) was ‘substantially motivated’ by the fact they were white; and (2) would not have been made without reference to plaintiffs’ race_” Uncontrovert-ed documentary evidence revealed that upon Mathis’s arrival as director of OMI, he found a division consisting of 32 police officers, 87 percent of whom were white males (three officers were black males and one was a female Hispanic). The only reason supported by the record for this statistical disparity is that former OMI director Maurer (a white officer with impeccable credentials) had staffed the unit with personnel he knew and trusted, and most of those people were white males. Neither party attempted to introduce evidence that the ratio of white to minority officers resulted from prior race discrimination. As the law now stands,8 without such proof Mathis would not have the legal right to institute an affirmative action program based solely on a “perceived” imbalance in the racial makeup of the department as Mathis had found it. See generally Janowiak v. Corporate City of South Bend, 836 F.2d 1034 (7th Cir.1987). Nevertheless, the evidence supports the inference (and the jury reasonably found) that Mathis took action to do something about the disparities based solely on his perception of the alleged under-representation of minorities in OMI personnel alone. The memo Mathis sent to William Ware, the mayor’s chief of staff, shortly after Mathis’s appointment provides this support: “The undersigned wishes to share with you the status of personnel assigned to OMI at the time I assumed command of this department, specifically its affirmative action posture. [Tabular listings followed, giving the numbers of ‘sworn personnel’ (police officers) and ‘civilian personnel,’ broken down in each instance as to ‘male’ and ‘female’ and within each of those categories, as ‘white,’ ‘black’ and ‘Hispanic.’] As this data indicates, affirmative action concepts do not exist in this department, either in its civilian or police components. Due to the obvious constraints in dealing with the civilian career service personnel, resolving this imbalance will be a long-range goal. However, in reference to the police personnel, this is being dealt with immediately. Further reports on this progress will follow.” After Mathis had made the transfers both out and in to the division, the unit more closely mirrored Chicago’s racial population — 836 — percentages. A second memo Mathis sent to Ware at the end of August 1984 revealed an overall decrease of white males from 28 to 15 (only two of those 15 had been with the office upon Mathis’s arrival). Conversely, the number of blacks went from three to 13 and Hispanics from one to four. Mathis did testify, albeit somewhat cryptically, that he instituted the transfers based on his desire to bring in people he knew and trusted, not discriminatory animus.9 But the jury properly rejected this testimony, finding as a matter of fact that his decisions were race-motivated. As this finding has support in the record, we must accept it as fact for the purposes of this appeal. This leads us to a possible second line of defense — affirmative action. As noted earlier, the defendants’ attorneys completely failed to label their defensive posture as such at any time before or during trial. But parts of Mathis’s testimony 10 and his memo of May 24, 1984, support the finding, and the jury specifically found (in response to interrogatory #4) that Mathis’s decision, to the extent it was race-motivated, “was perceived by Mathis as a means to correct what he considered a previously-existing racial imbalance within OMI.” While not cogently articulated, Mathis’s intent (as the jury found it) fits within the parameters of an affirmative action defense. Cf. McQuillen v. Wisconsin Education Ass’n Council, 830 F.2d 659, 666 (7th Cir.1987) (“If ... the employer claims that its hiring decision was based on the applicants’ relative qualifications, not an affirmative action program, a plaintiff must first establish that the employer was acting pursuant to an affirmative action plan before the plan’s validity is placed into issue.”) "... I am just saying in my opinion I didn’t look at it in that light. I looked at the positive way in terms of bringing more flexibility into the units and investigative responsibilities by bringing in people who were bilingual, who were female, who were white, who were black." Having said that much, we nevertheless conclude that the defense fails in this factual scenario. We begin our analysis of the defendants’ “affirmative action” defense with the fundamental principle that any distinctions between citizens based on race or ancestry are “by their very nature odious to a free people whose institutions are founded upon the doctrine of equality,” — 837 — Regents of University of California v. Bakke, 438 U.S. 265, 98 S.Ct. 2733, 2748-50, 57 L.Ed.2d 750 (1978), and thus are presumptively unconstitutional. See Personnel Admin. v. Feeney, 442 U.S. 256, 272, 99 S.Ct. 2282, 2292, 60 L.Ed.2d 870 (1979). Nevertheless, many cases on the question of equal protection hold that in order to correct the effects of past discrimination against minorities in this country, a narrowly tailored race-conscious affirmative action program will not, under limited circumstances (such as those explained below), violate the fourteenth amendment. See, e.g., United States v. Paradise, 480 U.S. 149, 107 S.Ct. 1053, 1064, 94 L.Ed.2d 203 (1987) (plurality opinion) (Fourteenth Amendment); United Steelworkers v. Weber, 443 U.S. 193, 208, 99 S.Ct. 2721, 2729, 61 L.Ed.2d 480 (1979) (Title VII). Recently, in Johnson v. Transportation Agency, 480 U.S. 616, 107 S.Ct. 1442, 94 L.Ed.2d 615 (1987), the Supreme Court clarified the manner in which the defense of a valid affirmative action program fits within the analytical framework for proof of unlawful racial discrimination set forth in McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973): “Once a plaintiff establishes a prima fa-cie case that race or sex has been taken into account in an employer’s employment decision, the burden shifts to the employer to articulate a non-discriminatory rationale for its decision. The existence of an affirmative plan provides such a rationale. If such a plan is articulated as the basis for the employer’s decision, the burden shifts to the plaintiff to prove that the employer’s justification is pre-textual and the plan is invalid. As a practical matter, of course, an employer will generally seek to avoid a charge of pretext by presenting evidence in support of its plan. That does not mean, ... that reliance on an affirmative action plan is to be treated as an affirmative defense requiring the employer to carry the burden of proving the validity of the plan. The burden of proving its invalidity remains on the plaintiff.” Johnson, 107 S.Ct. at 1449. As noted, the jury reasonably found that the City took race into account in making its transfer decisions. Further, the City has asserted, albeit inartfully, the existence of an affirmative action “plan” as the rationale for those decisions. Given the jury’s findings, in particular its response to special interrogatory # 4, we thus consider whether the City violated the equal protection clause of the fourteenth amendment when it instituted the wholesale transfer of sworn OMI personnel in response to a “perceived” statistical disparity between the percentage of minorities assigned to OMI and the percentage of minorities in some unspecified population group (either the Chicago police department or the City as a whole). (As a result of their waivering on this defense, the defendants fail to identify precisely which population group it allegedly compared with the OMI percentages — as we shall see, a critical problem.) Two recent Supreme Court decisions provide the backdrop for our analysis of the defendants’ “affirmative action” defense. In the first, Wygant v. Jackson Board of Education, 476 U.S. 267, 106 S.Ct. 1842, 90 L.Ed.2d 260 (1986), five justices in three separate opinions held that a race-preferential layoff provision in a collective bargaining agreement between the union representing school teachers in Jackson, Michigan, and the school board violated the equal protection clause of the fourteenth amendment. The provision, which the union and the board approved to help implement the board’s affirmative action plan, stated that: “[i]n the event that it becomes necessary to reduce the number of teachers through layoff from employment by the Board, teachers with the most seniority in the district shall be retained, except that at no time will there be a greater percentage of minority personnel laid off than the current percentage of minority personnel employed at the time of the layoff.” 106 S.Ct. at 1845. The Court was called upon to judge the constitutional validity of the provision when layoffs became necessary in 1974, and adherence to the collec — 838 — tive bargaining agreement would have resulted in the layoff of tenured, non-minority teachers while minority teachers on probationary status would be retained. Although the Wygant Court failed to reach a majority consensus on the constitutional standard applicable to all affirmative action plans, our Circuit has noted “that a ‘lowest common denominator majority position can be pieced together’ from the Wygant opinions.” Janowiak v. Corporate City of South Bend, 836 F.2d 1034, 1040 (7th Cir.1987) (quoting Britton v. South Bend Community School Corporation, 819 F.2d 766, 768 (7th Cir.1987)). At its broadest level of generality, the Wygant majority (and probably the entire court) agreed that (1) any affirmative action plan must be justified by a compelling government interest and (2) the means chosen by the government must be narrowly tailored to effectuate the plan’s purpose. Wygant, 106 S.Ct. at 1852. As we explained in Janowiak, the justices begin to part company on the question of when a “compelling interest” in remedying discrimination arises. Initially, Wygant’s three-justice plurality, authored by Justice Powell, firmly expressed its opinion that “[sjocietal discrimination, without more, is too amorphous a basis for imposing a racially classified remedy.” 106 S.Ct. at 1848. Justice Powell wrote that “[t]his court never has held that societal discrimination is sufficient to justify a racial classification. Rather, the Court has insisted upon some showing of pri- or discrimination by the government unit involved before allowing limited use of racial classifications in order to remedy such discrimination.” Id. at 1847 (emphasis added). Because of the Court’s consistent focus on “prior discrimination as the justification for, and the limitation on, a state’s adoption of race-based remedies,” Justice Powell states that “the relevant analysis in cases involving proof of discrimination by statistical disparity focuses on those disparities that demonstrate such prior governmental discrimination.” Thus, in Wygant, “the proper comparison for determining the existence of actual discrimination by the School Board was ‘between the racial composition of [the School’s] teaching staff and the racial composition of the qualified public school teacher population in the relevant labor market.’ ” Id. at 1847 (quoting Hazelwood School District v. United States, 433 U.S. 299, 307, 97 S.Ct. 2736, 2741, 53 L.Ed.2d 768 (1977)). As the School Board in Wy-gant proferred no such comparison to buttress its claim of past discrimination, the plurality found that the Board had not established a “compelling interest” for its affirmative action plan. 106 S.Ct. at 1849. In Janowiak we explained that a combination of this plurality ruling, Justice White’s concurrence, id. 106 S.Ct. at 1857, and Justice O’Connor’s concurrence, id. at 1852-57, produced a “lowest common denominator” holding that essentially squares with that of the plurality opinion. 836 F.2d at 1041. We stated: “For our purposes, the lowest common denominator holding of Wygant is that a statistical comparison upon which an affirmative action plan is based must compare the percentage of minorities in the employer’s work force with the percentage of minorities in the relevant qualified area labor pool before it can establish the predicate past discrimination required to justify an affirmative action remedy under the fourteenth amendment.” Id. at 1041-42. More recently, the Supreme Court in Johnson v. Transportation Agency, 480 U.S. 616, 107 S.Ct. 1442, 94 L.Ed.2d 615 (1987), applied a slightly less restrictive test for upholding the validity of an affirmative action plan. In finding a Santa Clara County, California, affirmative action plan permissible under Title VII, the Court held that “an employer seeking to justify the adoption of a plan need not point to its own prior discriminatory practices, nor even to evidence of an ‘arguable violation’ on its part.... Rather, it need point only to a ‘conspicuous [or manifest] ... imbalance in traditionally segregated job categories.’ ” 107 S.Ct. at 1451 (quoting Steel Workers v. Weber, 443 U.S. 193, 209, 212, 99 S.Ct. 2721, 2730, 2731, 61 L.Ed.2d 480 (1977)). — 839 — In determining that such a “conspicuous” imbalance exists as to justify a plan that takes race into account, Johnson confirmed the Wygant plurality’s holding: “A comparison of the percentage of minorities or women in the employer’s work force with the percentage in the area labor market or general population is appropriate in analyzing jobs that require no special expertise, ... or training programs designed to provide expertise, .... Where a job requires special training, however, the comparison should be with those in the labor force who possess the relevant qualifications.... The requirement that the ‘manifest imbalance’ relate to a ‘traditionally segregated job category’ provides assurance both that sex or race will be taken into account in a manner consistent with Title VII’s purpose of eliminating the effects of employment discrimination, and that the interests of those employees not benefiting from the plan will not be unduly infringed.” Id. 107 S.Ct. at 1452 (citations omitted). We noted in Janowiak, supra, that while the Johnson court never explicitly made the above comparison, it nevertheless found the manifest imbalance requirement satisfied where no woman held any of 238 skilled craft positions in the county. 836 F.2d at 1037-38. Thus, as Justice O’Con-nor pointed out, the Johnson majority implicitly found the manifest imbalance requirement satisfied by the existence of an “inexorable zero.” 107 S.Ct. at 1465 (O’Connor, J., concurring).11 Thus, for our purposes, where the City has proffered no evidence of an “inexorable zero” and absolutely no other evidence of past discrimination, any proferred statistical comparison must establish a manifest imbalance between the relevant qualified area labor pool (in this case those officers qualified to perform sensitive investigations of City corruption) and the OMI personnel. Mathis’s transfer program must fail under this analysis since the defense offered no such comparison. The jury’s broad finding — that Mathis’s transfer decisions were motivated by a desire to correct a previously existing racial imbalance “within OMI” — is of little help as the finding fails to delineate what population group (if any) the jury compared with the ratios of minority to non-minority OMI personnel. If the position of an OMI investigator requires that applicants possess minimum qualifications (which the position surely does), under Wygant and Johnson the City’s comparison would have to narrow its focus to those actually qualified for the position. The defendants not only fail to identify the focus of its comparison, but also fail to persuade us that Mathis even attempted to make any comparison whatsoever.12 (Mathis testified only that he — 840 — “perceived” an imbalance.) We hasten to emphasize that this omission is not the jury’s fault. The defendants’ failure to fully and properly articulate the factual predicate for their “plan” demonstrates that their race-based personnel decisions were not guided by a well-defined and narrowly tailored statistical analysis focusing on the relevant labor pool. “Absent a statistical comparison that focuses on the relevant qualified area labor pool, there is no assurance that race is being taken into account ‘in a manner consistent with Title VII’s [and the equal protection clause’s] purpose of eliminating the effects of employment discrimination.’ ” Janowiak, 836 F.2d at 1039 (quoting Johnson, 107 S.Ct. at 1452). The City’s “affirmative action” defense, therefore, fails to pass muster under the equal protection clause as a matter of law. The district judge thus properly upheld the jury’s findings of unjustified race discrimination, but it did so only with respect to those plaintiffs whose transfers were obviously “involuntary” (the plaintiffs Di Maggio and Cappitelli who were in the first wave of transfers) and with respect to those who could positively testify that they were replaced by a minority. Unfortunately, the court’s finely tuned, orchestrated and selective view of the evidence (unaided by a review of the trial transcript) fails to allow for an entirely reasonable possibility supported in the record — that the entire wave of transfers (taking place during a two-and-one-half-month period) was race-motivated and that this discrimination proximately caused every plaintiff’s transfer. As did the district judge, we group the plaintiffs into two groups for purposes of analysis. A. PLAINTIFFS CAPESIUS, GARTNER, AND WOJNAR The trial judge correctly noted that Capesius and Wojnar left OMI in mid-June 1984 at their own request, while Mathis affirmatively told Gartner that Gartner’s transfer would be forthcoming. Nevertheless, both Capesius and Wojnar testified that they spoke with their supervisors (plaintiffs Cappitelli and Di Maggio, whose transfers were race-motivated) who warned them that if they did not request a transfer out of OMI in the near future, they could be transferred without notice to any police department assignment. Based on these warnings and the events preceding the warnings, the court concluded that “it does not do violence to the evidence to say that it still rationally supports the conclusion Capesius and Wojnar (as well as Gartner) would in fact have been transferred out of OMI.” The court nevertheless entered JNOV against these plaintiffs because of its view that (1) “though a number of transfers made at Mathis’s instance were race-motivated, a substantial number were not,” and (2) “[e]ach of the plaintiffs under consideration here, by his conduct, rendered it impossible to say what would have caused his involuntary transfer had one been ordered.” After reviewing the entire record, we are convinced that in reaching this conclusion, the judge obviously and improperly substituted his view of the evidence for that of the jury. While the evidence might conceivably support the reasonable inference that some transfers were not race-motivated, and even that these plaintiffs were among that group, the evidence also supports reasonable inferences that cut the — 841 — other way.13 Indeed, the all-inclusiveness of Mathis’s memos to Ware, as well as the systematic nature of the transfers, weighs in favor of the conclusion that race played a part in all the transfers. Further, evidence that other plaintiffs were the subject of race discrimination during this relatively brief period of time (two and one-half months) is probative on Mathis’s intent as to these plaintiffs. Hunter v. Allis-Chalmers Corf., 797 F.2d 1417, 1423-24 (7th Cir.1986). Similarly, Mathis’s own testimony supports the inference that the entire series of transfers during the summer of 1984 was part of his overall plan to change the makeup of OMI personnel (and thus discriminatory motivation would equally affect each plaintiff’s transfer decision).14 Considering the entire body of evidence —including the supervisors’ warnings to these plaintiffs, the supervisors’ involuntary transfers, Mathis’s memos to Ware (which named all the plaintiffs), the pattern of involuntary transfers of white males, the speed of the reduction in white males, and Mathis’s clear race consciousness — a reasonable inference exists that these three plaintiffs’ transfers were the proximate result of Mathis’s race-based intent.15 Although there was little evidence of individual instances of behavior manifesting discriminatory motivation as to each particular plaintiff, we are convinced that the court’s failure to weigh and view the evidence evenhandedly and in its totality deprived the plaintiffs of all the reasonable inferences which could be drawn from the evidence, see Graefenhain v. Pabst Brewing Company, 827 F.2d 13, 19 (7th Cir.1987), including the inference that impermissible motivation played a part in the entire pattern of transfers. Thus, we conclude that the trial judge improperly substituted his choice over the jury’s rationale between two possible paths of inference reasonably supported in the evidence. B. PLAINTIFFS CYGNAR, FLANAGAN, O’DRISCOLL, RUBINO AND SHANAHAN Each of these five plaintiffs was part of an unwilling group of eight white officers which Mathis transferred on July 25, 1984. The court denied the defendants’ motion for JNOV as to Flanagan only because he testified that he was replaced by Tommy King, a black male officer, while the others could not so testify. The court found “But what that means is that the remaining four plaintiffs were transferred out, as part of a group of seven white officers, at the same time that six officers besides King were transferred in: four white males, one black male — 842 — and one black female (Mathis’s memo-randa having reflected, and his actions having demonstrated, he wanted to increase the number of female officers in the unit — only one had been there when he took over). That being true, it cannot be said the evidence, even with all reasonable inferences in each plaintiffs favor, can rationally support a finding that any of the Cygnar, O’Driscoll, Rubino, or Shanahan transfers was racially motivated. Judgment notwithstanding the verdict must be entered against each of them and in City’s favor.” (Emphasis added). (In its first order, the court entered JNOV against plaintiffs Bleke, Riggio and Murray on this same basis.) Again, we are persuaded that the trial court took an overly limited view of the evidence. We have previously noted that the replacement of minority employees by individuals of the same race does not preclude a finding of discriminatory intent. DeLesstine v. Fort Wayne State Hospital, 682 F.2d 130 (7th Cir.), cert. denied, 459 U.S. 1017, 103 S.Ct. 378, 74 L.Ed.2d 511 (1982). This is so because such “reasoning would foreclose a plaintiff from proving a prima facie case unless an employer discriminated not only against the plaintiff, but also against every so-called protected minority by hiring a so-called ‘non-protected’ person to fill that position.” Id. at 132. As the Supreme Court expressed: “A racially balanced work force cannot immunize an employer from liability for specific acts of discrimination.... ‘The company’s later changes in its hiring and promotion 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.’ ” Fumco Construction Corp. v. Waters, 438 U.S. 567 at 579, 98 S.Ct. 2943 at 2950-51, 57 L.Ed.2d 957 (1978) (citing Teamsters v. United States, 431 U.S. 324, 341-42, 97 S.Ct. 1843, 1857-58, 52 L.Ed.2d 396 (1977). While the trial judge properly recognized that the trier of fact can rationally treat the replacement of a white officer by a minority as evidence of discrimination, our cases hold that it is not, as the judge apparently concluded, necessary to demonstrate such replacement. Cf. Castaneda v. Partida, 430 U.S. 482, 499-500, 97 S.Ct. 1272, 1282-83, 51 L.Ed.2d 498 (1977) (holding that a woman may be guilty of discriminating against another woman). The court appears to theorize that the larger influx of black officers at the beginning of Mathis’s transfers demonstrates that as more blacks came in to the OMI unit, Mathis’s racial motivation disappeared. Even assuming the jury could properly have accepted this view of the evidence, the jury was entitled to and in fact did accept the contrary view that Mathis’s discriminatory motivation equally impacted all the plaintiffs and that the July 25 transfers were merely the last stage in his overall plan to change the racial makeup of OMI personnel. See, supra, n. 11. The jury’s verdict on the race discrimination claims must be sustained as to each plaintiff (and we conclude that a new trial is likewise unwarranted on these facts). C. MATHIS’S QUALIFIED IMMUNITY (RACE DISCRIMINATION) Mathis finally contends that he is immune from damages under the doctrine of qualified “good faith” immunity set forth in Harlow v. Fitzgerald, 457 U.S. 800, 102 S.Ct. 2727, 73 L.Ed.2d 396 (1982).16 — 843 — This issue is especially important here because only Mathis, not the City, is liable for punitive damages under § 1983, see City of Newport v. Fact Concerts, Inc., 453 U.S. 247, 271, 101 S.Ct. 2748, 2762, 69 L.Ed.2d 616 (1981), so if Mathis is legally immune from liability, the bulk of the plaintiffs’ damages award is unrecoverable. Under the doctrine of qualified immunity, government officials performing discretionary functions are shielded from liability for civil damages in an action brought under 42 U.S.C. § 1983 unless their conduct violated “clearly established statutory or constitutional rights of which a reasonable person would have known.” Harlow v. Fitzgerald, 457 U.S. at 818, 102 S.Ct. at 2738. See also Benson v. Allphin, 786 F.2d 268, 275 (7th Cir.1986). In Davis v. Scherer, 468 U.S. 183, 104 S.Ct. 3012, 82 L.Ed.2d 139 (1984), the United States Supreme Court stated that “whether an official may prevail in his qualified immunity defense depends upon the ‘objective reasonableness of his conduct as measured by reference to clearly established law.’ ” Id. 104 S.Ct. at 3018 (quoting Harlow, 457 U.S. at 818, 102 S.Ct. at 2738). The principle behind the doctrine is that “[i]f the law at that time was not clearly established, an official could not reasonably be expected to anticipate subsequent legal developments, nor could he fairly be said to know that the law forbade conduct not previously identified as unlawful.” Greenberg, 840 F.2d at 472 (quoting Harlow, 457 U.S. at 818, 102 S.Ct. at 2738). See also Rakovich, 850 F.2d at 1205-14. Applying these guidelines we must initially agree with the trial court that “[n]o area of constitutional law has been more unsettled than that gathered under the rubric of ‘affirmative action.’ ” As our earlier discussion of the defendants’ affirmative action defense demonstrates, fully two years after the events in this case the Supreme Court in Wygant remained badly splintered on a number of important issues in this area, including the standard applicable “to test the validity of the means chosen by [public officials] to accomplish ... race-conscious purposes.” 106 S.Ct. at 1849. Notwithstanding the continued development (and unclarity of) the law in this area, the plaintiffs urge in broad terms that Mathis should have known the principle that distinctions between citizens based on race or ancestry are “by their very nature odious to a free people whose institutions are founded upon the doctrine of equality.” Regents of University of California v. Bakke, 438 U.S. at 290-91, 98 S.Ct. at 2748. But the Supreme Court has emphasized that the test of “clearly established law” is not to be applied at such a broad level of generality. Otherwise “plaintiffs would be able to convert the rule of qualified immunity that our cases plainly establish into a rule of virtually unqualified liability simply by alleging violation of extremely abstract rights. Harlow would be transformed from a — 844 — guarantee of immunity into a rule of pleading. Such an approach, in sum, would destroy ‘the balance that our cases strike between the interests in vindication of citizens as constitutional rights and in public officials’ effective performance of their duties,’ by making it impossible for officials ‘reasonably [to] anticipate when their conduct may give rise to liability for damages.’ Davis, 468 U.S. at 195, 104 S.Ct. at 3019. It should not be surprising, therefore, that our cases establish that the right the official is alleged to have violated must have been ‘clearly established’ in a more particularized, and hence more relevant, sense: The contours of the right must be sufficiently clear that a reasonable official would understand that what he is doing violates that right. This is not to say that an official action is protected by qualified immunity unless the very action in question has been held unlawful, see Mitchell, 472 U.S. at 535, n. 12, 105 S.Ct. at 2820, n. 12; but it is to say that in the light of preexisting law the unlawfulness must be apparent.” Anderson v. Creighton, 483 U.S. 635, 107 S.Ct. 3034, 3039, 97 L.Ed.2d 523 (1987). At a more specific level of generality, case law in existence in 1984 held that in order to implement a valid affirmative action plan, there must exist: (1) some type of statistical disparity between the local labor force and the minority composition of the employer’s work force and (2) a time limit on the plan. Lehman v. Yellow Freight System, Inc., 651 F.2d 520 (7th Cir.1981). In these terms, it cannot be said that at the time Mathis instituted his transfer program, it had been “clearly established” that Mathis could not rely on the statistical disparity between minority personnel at OMI and the percentage of minorities in the Chicago area labor force. Further, Mathis’s implementation of his transfer policy was a one-time restructuring of OMI occurring over a three-month time period, thus meeting Lehman’s time limit requirement. Thus, although Mathis’s conduct may well violate today’s established constitutional law standards, nothing in Lehman (or other case law in existence at the time) clearly prevented Mathis from transferring the plaintiffs based upon their race to correct what he perceived to be statistical under-representation of minorities at OMI. As such, we believe Mathis is entitled to qualified immunity on the racial discrimination claims. POLITICAL DISCRIMINATION The Supreme Court has discussed the extent to which politically motivated discharges from employment infringe upon the first amendment rights of public employees in Elrod v. Burns, 427 U.S. 347, 96 S.Ct. 2673, 49 L.Ed.2d 547 (1976), and Branti v. Finkel, 445 U.S. 507, 100 S.Ct. 1287, 63 L.Ed.2d 574 (1980). Whether and under what circumstances the decision to transfer amounts to an unconstitutional impairment of a public employee’s right to freedom of speech and association under the first amendment based on the record is open to debate.17 As explained herein, we conclude that the court improperly entered JNOV against the plaintiffs with respect to the issue of the defendants’ motive on the political discrimination claims, and that a new trial is warranted on this issue. Thus, on remand, the court should also address the issue of whether a politically motivated transfer program states a cause of action under the circumstances presented here. Turning to the evidence of political motivation adduced at trial, the trial judge ruled that “[o]nly one bit of evidence could even arguably support the notion that any plaintiff’s transfer was ‘substantially motivated’ by his political affiliation: the fact that some plaintiffs had been purchasers of tickets to one or more former Mayor Byrne fundraiser dinners, plus some limited testimony that there was a — 845 — manila folder among the OMI files that contained some kind of list of purchasers (whether related to one or more Byrne functions is not clear).” Despite conceding the existence of this folder for the purposes of ruling on the post-trial motions, the court found that “[t]here was however no evidence (other than a highly attenuated inference) that the folder was ever known to — let alone seen by — Mathis [and] ... substantial doubt must attach to a finding based on such a weak foundation.” Whether Mathis knew of the contents of this folder (and thus those plaintiffs who supported Byrne in the general mayoral election) was disputed at trial. But contrary to the judge’s implicit conclusion that no reasonable jury could have found that Mathis knew of the folder’s existence, a review of the record reveals a chain of evidence from which the jury could very reasonably have so found. (This again points up the problems inherent in a trial judge’s second-guessing a jury’s factual determination without even ordering much less closely scrutinizing the entire transcript, even by the judge who heard all the evidence.) Initially, we note that Murray, OMI’s financial officer, testified that a list of contributors was maintained in a manila folder labeled “Jane Byrne Dinner” and that one of his co-workers (Ms. Danzig) brought the folder into his office with other financial records for his perusal. While the trial judge noted that “not all the plaintiffs even bought such tickets,” Officer Cappi-telli testified that he saw the list and that all the plaintiffs except Sergeant Di Mag-gio were named therein. Finally, Mathis himself testified that he in fact reviewed OMI’s financial records shortly after taking over. This chain of evidence, although not conclusive, obviously permits the inference that Mathis in fact reviewed the contents of the folder. While this evidence standing alone is hardly overwhelming on the ultimate question of Mathis’s intent, the trial court, in overturning the jury’s verdict, further failed to weigh the probative value of this evidence in the context of the rest of the record. In addition to evidence that Mathis knew of the plaintiffs’ political affiliation by means of the “Jane Byrne Dinner folder,” the record contains the following circumstantial indicia of discrimination based on party affiliation: (1) the pattern of transfers followed soon after a change in administration, cf. Meeks v. Grimes, 779 F.2d 417, 418 (7th Cir.1985) (upholding a district court’s finding of political motivation under the “clearly erroneous standard” of Fed.R.Civ.P. 52(e) based in part upon “the strong circumstantial inference arising from the close temporal proximity of the defendant’s assumption of office and the dismissals”); (2) the competence of the transferred employees (each plaintiff’s record contained high performance ratings; defendant Rice admitted that they all had “impeccable” credentials, see supra n. 1); and (3) statements which the plaintiffs’ supervisor made to the plaintiffs (e.g., plaintiff Cygnar testified that his new supervisor, Sergeant Nate Gibson, told him that even though “we didn't vote for the same people, we can work together”). Considered in its entirety, the combination of direct and circumstantial evidence provides a sufficient probative basis upon which a jury could reasonably reach a verdict in the plaintiffs’ favor on their political discrimination claim. Thus, the trial judge’s contrary ruling (granting the defendants’ motion for JNOV) cannot stand. One issue still remains. The judge conditionally granted the defendants’ alternative motion for a new trial in the event we reversed his order granting the defendants’ JNOV motion. As noted earlier, “[t]he authority to grant a new trial ... is confided almost entirely to the exercise of discretion on the part of the trial court ... [and thus] the grant or denial of a motion for a new trial is not subject to review by this court except upon exceptional circumstances showing a clear abuse of discretion.” General Foam, 695 F.2d at 288 (citations omitted). The judge’s grant of JNOV was error; still, we are not convinced that the court abused its discretion in ordering a new trial on this issue. While the evidence is sufficient to uphold the jury’s verdict, the verdict nevertheless is arguably “against the weight of the evi — 846 — dence” (the less stringent standard for the grant of a new trial); thus, we refuse to hold that the record articulates a clear abuse of discretion on the part of the trial court in granting the defendants’ alternative motion for a new trial. MATHIS’S QUALIFIED IMMUNITY (POLITICAL DISCRIMINATION) After Elrod and Branti (decided in 1980, four years before the transfers in issue here), it was well settled that political considerations could not properly motivate the firing of public employees unless “party affiliation is an appropriate requirement for effective performance of the public office involved [i.e., those employees in policy making or confidential positions].” Branti, 445 U.S. at 518, 100 S.Ct. at 1294. See also Danenberger v. Johnson, 821 F.2d 361, 364 (7th Cir.1987); Tomczak v. City of Chicago, 765 F.2d 633, 635 (7th Cir.1985).18 Nevertheless, Branti noted that the court was considering only the constitutionality of politically motivated discharges, 445 U.S. at 513 n. 7, 100 S.Ct. at 1292 n. 7, a limitation to which we later adhered in holding that the first amendment does not prohibit the use of political criteria in awarding public contracts. See LaFalce v. Houston, 712 F.2d 292 (7th Cir.1983). By 1984 only the Fourth Circuit had addressed the constitutionality of transferring a public employee from one position to another on the basis of his or her support for a particular political party. See Delong v. United States, 621 F.2d 618 (4th Cir.1980). And in Delong the court held that only in very limited circumstances amounting to “constructive discharge” does a politically motivated job transfer (such as is alleged here) create first amendment problems. 621 F.2d at 624. Thus, it was certainly not “clearly established” in the spring or summer of 1984 that transfers of the sort involved in this case — lateral transfers within a large organization that has frequent reassignments, involving no loss in rank or pay, and affording the transferees the opportunity to designate their preferences for new assignments — could not constitutionally take politics into consideration.19 Although other cases decided after the transfers in this case have taken positions contrary to Delong, see, e.g., Bennis v. Gable, 823 F.2d 723 (3d Cir.1987) (holding that any “disciplinary action” — in that case demotions —based upon political considerations violate the first amendment), Harlow holds that “[i]f the law at that time [of the official’s actions] was not clearly established, an official could not reasonably be expected to anticipate subsequent legal developments, nor could he fairly be said to ‘know’ that the law forbade conduct not previously identified as unlawful.” 457 U.S. at 818, 102 S.Ct. at 2738. In light of the case law that existed during the summer of 1984, we hold that there was no clearly established constitutional right that prohibited the defendants from transferring the plaintiffs based on political considerations. Mathis is therefore entitled to immunity on this issue.20 — 847 — DIRECTED VERDICT AS TO MAYOR WASHINGTON The plaintiffs maintain that the district court erred in granting the defendants’ motion for a directed verdict as to former Mayor Washington. Although the mayor may have indirectly (through Ware) approved Mathis’s transfer program, there is no doctrine of superior’s liability in § 1983 actions; instead, the official must actually have participated in the constitutional wrongdoing. See, e.g., Ustrak v. Fairman, 781 F.2d 573, 575 (7th Cir.1986); Crowder v. Lash, 687 F.2d 996, 1005 (7th Cir.1982). Here there is no evidence that Mayor Washington did. He may have failed to stop the transfers; but as we explained in Soderbeck v. Burnett County, 752 F.2d at 293: “Failure to take corrective action cannot in and of itself violate § 1983. Otherwise the action of an inferior officer would automatically be attributed up the line to his highest superior and thence to the local government, since the misconduct of policy making officials ... is attributed to the government. This ladder of liability would be inconsistent with the distinction created by Monell [v. New York City Department of Social Services, 436 U.S. 658, 98 S.Ct. 2018, 56 L.Ed. 2d 611 (1978)] between the direct and derivative liability of local government, and with the many lower-court cases that reject the notion that supervisors are strictly liable for their subordinates’ violations of § 1983. See, e.g., Schultz v. Baumgart, 738 F.2d 231, 238-39 (7th Cir. 1984); Wolf-Lillie v. Sonquist, 699 F.2d 864, 869 (7th Cir.1983).” We hold that the trial court properly granted the defendants’ motion for directed verdict with respect to former Mayor Washington. DAMAGES As noted, Mathis’s entitlement to qualified immunity vitiates the jury’s award of punitive damages. That leaves us with the question of the propriety of the jury’s compensatory damage award in thf amount of $55,000 per plaintiff. The district court found the award excessive to compensate the plaintiffs for essentially “intangible” injuries, and thus granted the defendants’ motion for a new trial on the damage issue, unless — at his option — each plaintiff accepted a remittitur to $15,000. The federal standard for reviewing the dollar amount of a jury award has been the subject of some change in recent years. We have traditionally noted that only if the jury damage award is “monstrously excessive,” a product of passion and prejudice, or if there is no rational connection between it and the evidence, may the trial court disturb it. See Matter of Innovative Construction Systems, 793 F.2d 875, 887 (7th Cir.1986) (collecting cases). “Underlying this deference to a jury’s assessment of damages is the acknowledgment that ‘the actual measure of damages is an exercise in fact-finding,’ and any rule on the scope of the judge’s control of the jury has Seventh Amendment implications.” Id. at 887-88 (quoting Whitley v. Seibel, 676 F.2d 245, 252 (7th Cir.), cert. denied, 459 U.S. 942, 103 S.Ct. 254, 74 L.Ed.2d 198 (1982)). Thus, while the trial judge’s determination of the propriety of a jury’s damage award warrants substantial deference (because of — 848 — his or her superior position to assess the credibility of witnesses and the like), “a somewhat more exacting standard of appellate review should apply” where the court grants, rather than denies, a new trial on damages. Id. at 888. See also Knapp v. Whitaker, 757 F.2d 827, 847 (7th Cir.1985). In addition to this traditional formulation of the standard of review, recently “this court added an additional element to the equation where the case under review is but one of a series of similar cases that establish a trend in damage awards,” i.e., compatibility among such awards. Joan W. v. City of Chicago, 771 F.2d 1020, 1023-24 (7th Cir.1985); see also Levka v. City of Chicago, 748 F.2d 421, 425 (7th Cir.1984) (“One factor we must consider in determining whether to set aside an award is whether the award is out of line compared to other awards in similar cases”); Bailey v. Andrews, 811 F.2d 366, 374-76 (7th Cir.1987). After a thorough examination of the record, we conclude that the district court properly ordered a new trial on the issue of damages with the option of remittitur to $15,000. Initially, the trial court correctly and reasonably found that the jury’s identical $55,000 award to each plaintiff — despite sharp variances among them in any asserted economic harm (some plaintiffs testified that they had suffered limited economic harm such as the loss of secondary job opportunities, others failed to testify as to economic harm, and some expressly denied such harm) — means that the entire award must have been attributed to “intangible” harm such as damage to reputation, humiliation, embarrassment, and emotional distress (which all of the plaintiffs testified they had suffered). Unquestionably, damages may be awarded for such items of intangible injury resulting from a violation of constitutional rights. See Carey v. Piphus, 435 U.S. 247, 264, 98 S.Ct. 1042, 1052, 55 L.Ed.2d 252 (1978). Still, the compensatory damage award must bear a reasonable relation to actual injury sustained; it may not be punitive in nature.21 Last year, in Webb v. City of Chester, 813 F.2d 824, 836 (7th Cir.1987), we noted that “[o]ur review of other cases in which plaintiffs were fired from their jobs” (lower damages would be expected with respect to transfers) revealed awards ranging from a “low of $500 to a high of over $50,000.” 813 F.2d at 836-37 nn. 3, 4. On that basis, the court upheld a jury award of $20,250 for “embarrassment and humiliation” resulting from a violation of the plaintiff’s constitutional rights. Id. at 837. We find Webb’s result instructive. The jury’s award of $55,000 in this case — significantly higher than any award approved in the context of unconstitutional firings (not transfers as in this case) as surveyed in Webb — is clearly excessive for a case involving no loss of position within the Chicago police department. As the judge’s award of a new trial on the issue of damages with the option of remittitur to $15,-000 is in line with other awards in similar cases, we approve the court’s disposition of the damages issue. CONCLUSION Accordingly, the decision of the district court is affirmed in part, reversed in part) and remanded for further proceedings consistent with this opinion. Circuit Rule 36 shall apply on remand. Footnotes [1] Maurer selected OMI's "sworn investigators” from the ranks of the Chicago police department. In choosing his officers, he looked for those who had investigative skills and experience, strong ability in report writing, strong educational background, and demonstrated honesty, integrity and loyalty to their jobs. Maurer selected the appellants for these positions because they met those criteria. Of the 13 appellants, ten had served as detectives with the Chicago police department. All had an average of 15 years’ investigative experience with the department. All of the appellants’ performance ratings were very high, and none had a disciplinary history. Many had special talents and skills which were needed in OMI’s work. Four of the appellants had post-graduate degrees, and six had bachelor's degrees. All of the appellants had received special recognition for their work with the Chicago police department prior to their assignment to OMI. In Maurer’s opinion, the appellants were superior officers; defendants Rice and Mathis acknowledged that their records were "impeccable.” During the time they served under Maurer, the appellants had taken part in successful and highly praised investigations into official corruption, theft and bribery involving the Chicago Housing Authority, the Department of Revenue, the Animal Control Unit, and Duncan Parking Meters. [2] Officer Riggio testified that he requested a transfer because his partner, Officer Cygnar, was part of the July 25 transfer order, and he felt that Mathis had lied to him about their job security in the OMI detective positions. Officer Bleke stated that he left after his new supervisor informed Bleke that he would probably be going. Plaintiff Murray, OMI’s financial officer, — 832 — testified that he requested a transfer after another sergeant started taking over his duties and after Mathis refused to meet with him regarding OMI financial matters. [3] As with their answer, the defendants’ motion for a directed verdict broadly (and without reference to affirmative action) raised the affirmative defense of qualified immunity. In a brief, one-sentence paragraph: “This honorable court should direct a verdict or in the alternative enter judgment in favor of the defendants Washington, Rice and Mathis in their personal capacities and against the plaintiffs for the reason that the evidence clearly demonstrated that at all times they acted in accordance with the law and in good faith and with the reasonable belief that they were so acting. Harlow v. Fitzgerald, 447 [457] U.S. 800 [102 S.Ct. 2727, 73 L.Ed.2d 396] (1982).” [4] Curiously, the transcript of the court’s instruction conference makes no mention of special interrogatory no. 4. At oral argument, counsel for the plaintiff represented that the trial judge "put together” the interrogatory. Because of the absence of any discussion regarding this interrogatory in the record, we are unable to ascertain whether the parties objected to its submission to the jury. [5] The trial court granted the defendants’ motions for a directed verdict on the plaintiffs’ due process claim and all the plaintiffs’ claims against the defendant Washington as well. Plaintiffs appeal only the latter ruling. Further, the plaintiffs voluntarily dismissed their Title VII claim following the jury's rendering of its verdict. Hence, only the court’s ruling on the plaintiffs’ race and political discrimination claims are before this court on appeal. [6] The court surprisingly refused to withhold final ruling on the defendants’ post-trial motions pending the preparation of a trial transcript despite counsel's representation (during a hearing on the plaintiff’s motion for certification under Rule 54(b) held more than three weeks before the entry of the court’s second order on the defendant’s post-trial motions) that one of the attorneys would soon be ordering the record. Apparently concerned that the preparation of a transcript would unduly delay final disposition of the case, the court stated: "I’d rather have them [the parties’ submissions] from your notes. I don’t want to defer this. You know, it’s been long enough anyway, and the idea of awaiting the preparation of a transcript I don’t think serves anybody. I’d much rather have the prompt responses I had asked for from you.” [7] Although issued on the final page of the first of the court’s two orders on the defendants’ post-trial motions, this ruling apparently covers both orders. We therefore construe the ruling as such. [8] An important qualification, as we shall see in the context of analyzing the qualified immunity question. [9] For instance, in response to questions about whether the plaintiffs’ race played a part in his transfer decisions, Mathis stated: [10] Mathis’s testimony with respect to affirmative action is as follows: Q. Isn’t it a fact the only thing that you put down there is that there is an affirmative action posture? Isn’t that the only thing you referred to is that you have said that this is a problem — affirmative action problem? A. No, I mentioned that but that wasn’t the intent of this memorandum. Q. Okay. Well, show me what the intent of the memorandum is in the words of the memorandum? A. I am just saying the intents of the memorandum are just the stats that are shown here, period, in terms of stating to Mr. Ware just what I found in terms of personnel staff, period. Q. Okay. You thought you would send these statistics up to Mr. Ware because he might just be interested in reading these statistics, is that correct? A. Certainly. Q. Okay. And you made no mention there of how this imbalance decree created a problem for— whether or not OMI was going to run well as an agency? A. No, I didn’t. Q. Okay. But you did talk about affirmative action; isn't that correct? A. Yes. Q. And you talked about it on Page 1 and you talked about it on Page 2? Isn’t that correct? A. Yes, yes. Q. In fact you underlined “do not exist,” isn’t that correct? A. Yes, concepts. Q. Okay. Is that an indication of strong feeling on your part, the fact that you underlined that; was that something that you wanted to emphasize? A. No, I just wanted to make sure that he understood what I was saying, that the concepts of affirmative action in terms personnel extent [sic] did not exist there, that's all. * [11] While finding the existence of a manifest imbalance, the majority in Johnson also took pains to emphasize that the Santa Clara County plan did not unnecessarily trammel the rights of non-minorities. See Janowiak, 836 F.2d at 1038. For instance, the Court noted that an affirmative action plan may not amount to mere “blind hiring” by the numbers: in this vein, the Court elaborated as follows on the statistical analysis necessary to determine whether a manifest imbalance exists in a particular job category: "[H]ad the plan simply calculated imbalances in all categories according to the proportion of women in the area labor pool, and then directed that hiring be governed solely by those figures, its validity fairly could be called into question. This is because analysis of a more specialized labor pool normally is necessary in determining under-representation in some positions. If a plan failed to take distinctions in qualifications into account in providing guidance for actual employment decisions, it would dictate mere blind hiring by the numbers, for it would hold supervisors to ‘achievement of a particular percentage of minority employment or membership ... regardless of circumstances such as economic conditions or the number of qualified minority applicants.’ ” Id. at 1454 (quoting Sheet Metal Workers v. EEOC, 478 U.S. 421, 106 S.Ct. 3019, 92 L.Ed.2d 344 (1986) (O’Connor, J., concurring in part, dissenting in part)). "Instead of authorizing such blind hiring, the Santa Clara County plan 'expressly directed that numerous factors be taken into account in making hiring decisions, including specifically the qualifications of female applicants for particular jobs.’ ” Janowiak, 836 F.2d at 1038 (quoting Johnson, 107 S.Ct. at 1455). [12] The defendants imply that Mathis did not have to demonstrate he made such a comparison given that "[t]he long history of the Chicago police department’s discrimination against blacks and other minorities is a matter of judicial record.” See United States v. City of Chicago, 549 F.2d 415 (7th Cir.1977). This argument — 840 — rests on the false assumption that past discrimination in the entire Chicago police department means past discrimination within OMI. The defendants had ample opportunity at trial to attempt to establish a pattern of discrimination at OMI but, as noted, failed to do so. In our view, Johnson and Wygant require a more particularized showing of past discrimination (by statistical comparison or otherwise) before a municipality may employ a race-preferential plan. Cf. Watson v. Fort Worth Bank & Trust, — U.S.-, 108 S.Ct. 2777, 2789, 101 L.Ed.2d 827 (1988) ("the plaintiff [in a disparate impact case] must offer statistical evidence of a kind and degree sufficient to show that the practice in question has caused the exclusion of applicants for jobs or promotions because of their — 841 — membership in a particular group.”); Shidaker v. Tisch, 833 F.2d 627, 631 (7th Cir.1987). [13] In the context of a criminal appeal the Supreme Court has cautioned appellate judges to avoid judicial speculation: “[I]t is not the appellate court’s function to determine guilt or innocence. Nor is it to speculate upon probable reconviction and decide according to how the speculation comes out. Appellate judges cannot escape such impressions. But they may not make them sole criteria for reversal or affirmance. Those judgments are exclusively for the jury, given always the necessary minimum evidence legally sufficient to sustain the conviction unaffected by the error.” Kotteakos v. United States, 328 U.S. 750, 763-64, 66 S.Ct. 1239, 1247, 90 L.Ed. 1557 (1946). [14] During this exchange as part of the plaintiffs’ case-in-chief, Mathis testified as follows: "Q. Now — so you went back to the police — to OMI and you started rectifying the imbalance as you perceived it, is that correct? A: I went back to my office and proceeded with my daily duties on that particular day whatever they might be — might have been, that is. Q: Did you make any changes after this meeting with Mr. Ware? A: I made changes following this particular date up until August I believe. I was doing interviewing of quite a few personnel.” [15] The trial judge stated that this result can only be reached if the jury engages in speculation as to what Mathis’s intent would have been had he in fact decided to transfer these plaintiffs. While the line between reasonable inference and speculation is often not clearly defined, we are convinced that the jury did not have to speculate that Mathis’s discriminatory motivation extended to all the plaintiffs (as noted, the evidence supports the inference that Mathis instituted a uniform pattern of transfers based upon the results of his survey of OMI personnel). [16] The plaintiffs insist that Mathis waived the defense of qualified immunity through his failure to proffer a jury instruction on the issue or move for summary judgment on that basis. Initially, the defendants’ failure to proffer a jury instruction on the issue is immaterial. As Mitchell v. Forsyth, 472 U.S. 511, 527-28, 105 S.Ct. 2806, 2816, 86 L.Ed.2d 411 (1985) makes clear, qualified immunity means immunity from standing trial, and hence the issue is one of law for the court’s resolution. Greenberg v. Kmetko, 840 F.2d 467, 472 (7th Cir.1988). Thus, the ultimate issue of immunity is not properly the subject of an instruction to the jury. Cf. Rakovich v. Wade, 850 F.2d 1180, 1201-02 (7th Cir.1988). The point is well taken, however, that the defendants should have filed a motion for summary judgment on the qualified immunity issue. In Walsh v. Mellas, 837 F.2d 789, 799-800 n, 7 (7th Cir.1988), we noted that Harlow itself stated that the most appropriate time — 843 — to raise the qualified immunity issue is in a motion for summary judgment filed before allowing discovery. And recently, in Anderson v. Creighton, 483 U.S. 635, 107 S.Ct. 3034, 97 L.Ed.2d 523 (1987), the Supreme Court again emphasized what has always been the central purpose of the qualified immunity defense: "To protect public officials from the “broad ranging discovery’ that can be particularly disruptive of effective government. For this reason, we have emphasized that qualified immunity questions should be resolved at the earliest possible stage of a litigation.” 107 S.Ct. at 3042 n. 6. While the defendants’ failure to diligently pursue the qualified immunity issue on a pretrial motion has frustrated these important goals, we decline to hold that the defendants waived the issue where they filed a motion for a directed verdict on this basis, the trial court premised its ultimate decision as toj punitive damages on the issue, and such damages make up the bulk of the jury’s overall $4.29 million verdict. Under our specific facts, we believe that “exceptional circumstances [exist] where justice demands more flexibility” in the application of the waiver rule. Johnson v. Artim Transportation System Inc., 826 F.2d 538, 547-48 (7th Cir.1987). Although the result in this scenario may well be different under other circumstances, we decline to hold that Mathis’s failure to present the issue in a summary judgment motion waived his right to assert the defense in a later JNOV motion or on appeal. Cf. Rakovich v. Wade, 850 F.2d 1180, 1199-1200 (7th Cir.1988) (issue of qualified immunity not waived despite being raised for the first time in a pre-trial brief, followed by a directed verdict on that basis). [17] Compare Delong v. United States, 621 F.2d 618 (4th Cir.1980) (only those patronage practices that "can be determined to be the substantial equivalent of dismissal” impermissibly burden a public employee's first amendment rights) with Bennis v. Gable, 823 F.2d 723 (3d Cir.1987) (any "disciplinary action” against a public employee based upon political affiliation violates the first amendment). [18] Notwithstanding Branti and Elrod, we noted in Soderbeck v. Burnett County, 752 F.2d 285, 291 (7th Cir.1985), that "the discharge of public employees on political grounds is not yet regarded in the light of something contrary to natural law; nor is it widely known in non-legal circles to be a federal constitutional tort or any other sort of tort." [19] The plaintiffs also urge that the consent decree entered into by the City of Chicago in Shakman v. Democratic Organization, 356 F.Supp. 1241 (N.D.Ill.1972), should have placed Mathis on notice that his "personnel decisions" could not be politically motivated. The plaintiffs’ reliance on the so-called “Shakman decree” is misplaced in two respects. Initially, until the district court decided Herron v. City of Chicago, 591 F.Supp. 1565 (N.D.Ill.1984), decided a month after Mathis’s transfers were completed (ruling that the decree covered certain promotions), it was unclear as to what employment actions beyond discharge the decree covered. See Herron, 619 F.Supp. 767, 769 (N.D.Ill.1985). More importantly, it is not clear that the OMI investigator positions are not "exempt” from the Shakman decree based upon the confidential nature of the positions. [20] Further, we agree with the district court that in 1984 (or even now) it was not clearly established that the transferred OMI personnel (an elite group designed to investigate City corruption) were not "confidential’’ employees for whom "party affiliation is an appropriate requirement for effective performance of the pub847 lie office involved.” Branti, 445 U.S. at 518, 100 S.Ct. at 1294. Indeed, this language from Meeks v. Grimes, 779 F.2d 417, 420 (7th Cir.1985), makes for a strong argument that the plaintiffs were "confidential” employees within the meaning of Branti: "Thus our attention turns to the second type of office for which politics is a constitutionally permissible job criterion; these positions can be generally described as ‘confidential.’ This is a catch-all phrase that encompasses those government employees who, while not decision makers, are in close contact with policy makers and the highly confidential communications or records affecting decisions. See Soderbeck, 752 F.2d at 288 (“You cannot run a government with officials who are forced to keep political enemies as their confidential secretary.'). The basis for exempting this type of employee is that political antipathy can serve as a decent proxy for a lack of trust and loyalty where the employee’s responsibilities include a duty to shield the decision making process from the outside world. The possibility of ‘leaks’ from employees with access to sensitive information is a constant threat to any unit of government.” Mathis is entitled to immunity on this basis as well. [21] The plaintiffs surprisingly suggest that the court could have sustained the $55,000 damage award by splitting it into two parts and awarding $27,500 for each of the two constitutional violations (political and racial discrimination). Memphis Community School Dist. v. Stachura, 477 U.S. 299, 106 S.Ct. 2537, 2545, 91 L.Ed.2d 249 (1986), puts this clever (but disingenuous) argument to rest. The case holds that a jury may not award damages based on the abstract "value" of a constitutional right; it must award damages suffered (including intangible injury). To “split” the actual damage award would be to award damages based on the supposed value of the right itself, something Stachura says we may not do.