Summary
finding classwide pro rata reductions appropriate where thirty-five officers were eligible for two promotions
Summary of this case from Munoz v. Bd. of Educ. of City Sch. Dist. of N.Y.Opinion
No. 98-4626
DECIDED November 17, 1999
Albertine B. Smith, Myrna D. Bricker, Miami, FL, Dennis J. Dimsey, Marie K. McElderry, U.S. Dept. of Justice, Civil Rights Div., Washington, DC, for Plaintiff-Appellant.
Robert David Klausner, Hollywood, FL, for Defendant-Appellee.
Bonnie Joyce Losak-Jimenez, Joseph W. Beasley, Kelly, Black, Black, Byrne Beasley, Miami, FL, Stephen H. Cypen, Cypen Cypen, Miami Beach, FL, for Intervenor.
Appeals from the United States District Court for the Southern District of Florida.
This appeal is the latest chapter in a lengthy litigation saga over a 1977 nondiscrimination consent decree between the United States of America, the City of Miami, and the Fraternal Order of Police, Lodge No. 20. In this iteration, the Fraternal Order of Police ("Appellee"), on behalf of a class of white and Hispanic police officers, brought two civil contempt actions against the City of Miami for "reverse" race discrimination in its 1992 police officer promotion practices. The district court found that the City had discriminated on the basis of race in its "special certification" of several minority promotion candidates, resulting in the unlawful promotion of one black lieutenant and one black sergeant. As a result, the district court held the City in civil contempt of the 1977 consent decree. The district court then awarded broad "make-whole" relief to all "adversely affected" police officers, as if each of these officers actually would have received one of the two promotions in 1992.
On appeal, the United States of America and the City of Miami ("Appellants") do not challenge the district court's finding of race discrimination or civil contempt. The sole issue before us is whether the district court abused its discretion in fashioning broad "make-whole" relief for the entire officer class. After a thorough review of the record and the parties' briefs, we conclude that the district court's chosen remedy was excessive and that it should have divided the monetary value of the two promotions on a pro rata basis amongst the class of eligible candidates. We therefore vacate, in part, the judgment filed on March 13, 1998, and remand with instructions for its modification consistent with this opinion.
I.
The facts of this prolonged case began in 1975 when the United States of America sued the City of Miami, various City officials, and several police officer unions for discriminatory employment practices adversely affecting black, Hispanic, and female individuals in police hiring and promotion policies in violation of Title VII of the Civil Rights Act of 1964, the Fourteenth Amendment, and 42 U.S.C. §§ 1981 and 1983. The United States and the City agreed to settle the case, and the district court approved a consent decree in 1977 over the objections of a police union, the Fraternal Order of Police ("FOP"). The consent decree required the City to establish promotional "goals" for protected minority groups.
The decree contains the following relevant provisions:
1. The defendant City of Miami, its officials, agents . . . are permanently enjoined and restrained from engaging in any act or practice which has the purpose or effect of unlawfully discriminating against any employee of, or any applicant, or potential applicant for employment with the City of Miami because of such individual's race, color, sex, or national origin. Specifically, the City shall not refuse to hire, promote, upgrade, train, or assign any individual, discharge any individual, or otherwise discriminate against any individual as an employee or applicant for employment with respect to compensation, terms, conditions, or privileges of employment because of race, color, sex, or national origin.
. . .
5. Goals
In order to eliminate the effects of past discriminatory practices against blacks, Latins, and women, the City shall adopt and seek to achieve as its long term goal the participation at all levels throughout its workforce of blacks, Latins, and women approximating their respective proportions in the City's labor force. . . . The purpose of this goal is to eliminate the substantial under-representation and uneven distribution of blacks, Latins, and women throughout the City's workforce. . . .
(b) Promotions
Subject to the availability of qualified applicants, promotional goals shall be established for minorities. . . .
At the time, police promotion decisions were governed by a civil service rule, Ordinance No. 6945, known as the "Rule of One." The rule required the City to hire and promote certified applicants rigorously based on their designated rank order on the eligible candidate register. This rank order was determined solely by the results of the civil service promotional exam. On April 17, 1978, the Justice Department wrote a letter to the Miami City Manager outlining how the City's "Rule of One" adversely affected minority employees in violation of the consent decree. In July 1979, the City amended its civil service rules, adopting Ordinance No. 8977 which, among other things, substituted a "Rule of Eight" for the previous "Rule of One."
Under the new rule, eight candidates must be certified for each vacant position by the City Director of Personal Management. The first five candidates must be chosen in rank order from the results of the promotional exam. However, the Director also has the discretion, as affirmative action needs dictate, to certify three minority officers by exam rank order. For each additional vacancy, two candidates are to be added to the certified candidate pool: the officer with the next highest overall test score and the minority officer with the next highest test score. In addition, the City also passed a "special certification" rule, Rule 8.7, that allows the Director to certify up to three additional candidates if special requirements of sex or domicile are involved, or additional "special qualifications" are required. All certified candidates must score high enough on the promotional exam to satisfy the City's eligibility requirement. Once a list of eligible candidates has been certified, the Miami Police Chief conducts interviews with each certified candidate. The Chief then chooses his promotions based on the interview and other subjective factors. A candidate's written exam score plays no role in this final decision-making process.
In 1981, we finally heard the FOP's appeal of the consent decree. We approved the decree's provisions pertaining to the City and the United States, but adjudged those provisions pertaining to the FOP invalid because the FOP had not consented to the decree. See United States v. City of Miami, 614 F.2d 1322 (5th Cir. 1980), vacated in part, 664 F.2d 435, 438-39 (5th Cir. 1981) (en banc). On remand, the district court entered a consent order on April 4, 1983, to which the FOP voluntarily consented, that maintained in full force and effect the provisions of the earlier consent decree including the "Rule of Eight." The City still employs the "Rule of Eight" for all police hires and promotions.
In 1983, the Miami Association of Firefighters, Local 587 challenged the City's implementation of the "Rule of Eight." The district court ruled that the union had signed the consent decree and therefore was bound by its terms, particularly including the "Rule of Eight." The district court also determined that the "Rule of Eight" was a legitimate tool to remedy the effects of past City discrimination. Local 587 appealed, but the appeal was dismissed for procedural reasons. See United States v. City of Miami, 2 F.3d 1497, 1500-01 (11th Cir. 1993). Because this litigation involves only the Police and not the Fire department, we limit our discussion of the history of all relevant litigation involving the Fire Department to the extent necessary to understanding the issues raised in this appeal.
On March 24, 1992, a promotional register for the rank of police lieutenant was posted by Director Angela Bellamy based on the results of the written exam. On April 8, Miami Police Chief Calvin Ross requested a certification list to fill sixteen police lieutenant vacancies. That same day, the Chief also requested in a memo that additional black candidates be "specially certified" for the position in accordance with Rule 8.7 to provide "positive role models and ethnic recognition." At the time, two of twenty-nine lieutenants (roughly seven percent) in the Miami police department were black. On May 18, Director Bellamy ordered that a portion of the promotional exam be re-administered due to errors in the original exam. A new exam was given on May 28 and June 1. On June 19, a new eligible register was established. On July 21, 1992, Police Chief Ross again requested Director Bellamy to issue a certification list to fill the sixteen vacancies. In a same-day memo, Ross also re-requested that additional black candidates be "specially certified" in accordance with Rule 8.7. The memo explained that black supervisors were needed to "understand and speak" non-traditional English. It also noted that black supervisors were needed because many drug enforcement operations occurred in black neighborhoods, requiring supervisors who could "blend" into the environs and make "command-level decisions."
On July 23, three black male candidates were certified under the "special qualifications" requirement — Roberson Brown, Jr., Bobbie J. Meeks, and Gerald L. Darling. Each of these candidates, based on their exam scores, would have been eligible under the Rule of Eight for certification even if no "special certification" had taken place. All three received a promotion to lieutenant. However, as a result of the "special certification," three additional minority candidates were made eligible for a lieutenant promotion: Alphonso Erving, Javier Mayor, and Melvin Uptgrow. One of these candidates, Erving, actually received a promotion. Erving would not have been eligible for promotion if the "special certification" had not occurred — i.e., his test score was too low to be included in the normal certification pool of 38 candidates but was high enough to be included once the pool was expanded to 41 candidates based on the "special certification." In the end, four white males, six Hispanic males, four black males (including Erving), one Hispanic female, and one white female were promoted to lieutenant.
On November 25, 1992, Chief Ross requested Director Bellamy to add several Creole speakers through "special certification" to the eligible candidate register for five available police sergeant positions. Two Creole-speaking candidates were then added to an original candidate list of 16 persons: Gary Eugene, a black male ranked 107th on the eligible register and Mario Roman, a Hispanic male ranked 77th on the eligible register. Eugene's employment application recorded his Creole skills as advanced while Roman's application recorded his Creole skills as beginner. The police chief selected Eugene over Roman for one of the sergeant positions.
In 1993, the FOP filed two civil contempt motions against the City for consent decree violations stemming from these police lieutenant and sergeant promotions. The FOP argued that the City had employed its "special certification" procedure to unlawfully select officer candidates solely on the basis of race. The first motion alleged that the City's "special certification" of several black officers, in response to the Police Chief's request for black lieutenants to supervise undercover operations in predominantly black neighborhoods, was a pretext for promoting lieutenants on the basis of race. The second motion claimed that the City's "special certification" of two minority sergeant candidates, in response to the Police Chief's request for Creole-speaking sergeants, was a pretext for promoting sergeants on the basis of race or ethnicity. Both parties agreed to have the motions resolved on the case record and submitted a joint statement of undisputed facts on March 14, 1997. The United States requested an evidentiary hearing on the merits of the motions but the request was denied.
On March 13, 1998, the district court granted the FOP's motions. In so ruling, the court interpreted the consent decree to prohibit the designation of race as a "special qualification" for promotion. The court then dismissed both City "special certification" reasons as being pretextual. With respect to the lieutenant promotions, the court specifically found that since lieutenants serve a supervisory rather than a field capacity, there was no legitimate need to select candidates on the basis of their racial ability to "blend" into target communities. The district court also found that the City had made no showing of how lieutenant command functions required officers of a particular race. As a result, the court determined that the City improperly certified three black candidates out of forty-one total candidates, of which, one improperly received a promotion.
With respect to the sergeant promotions, the district court found that ethnicity rather than Creole was the predominant "special qualification" for the "special certification" of two sergeant candidates. In its findings, the district court determined that police sergeants conducted their official responsibilities in English rather than Creole, and that the City had not tested sergeant candidates for language skills nor mentioned language in its advertisements for the sergeant promotions. The district court finally concluded that City civil service rules allowed for the temporary appointment of Creole-speaking sergeants if an emergency situation required it. One sergeant candidate, Gary Eugene, received a promotion after being improperly certified under the City's "special certification" procedures.
To remedy these certification violations of the consent decree and the concomitant promotion of one lieutenant and one sergeant, the district court ordered extensive "make-whole" relief for all police employees "adversely affected." This class includes the twenty-three lieutenant and twelve sergeant candidates eligible for promotion who were bypassed in favor of the two candidates improperly certified through the City's "special certification" rule. All class members were awarded full backpay, retroactive seniority, a fifteen thousand dollar lump-sum pension payment, and a one rank promotion from their current positions. This relief was ordered despite the fact that only one lieutenant and one sergeant candidate obtained a promotion in 1992 as a result of the City's unlawful certification procedures.
Because the district court's order never precisely defines the "adversely affected" officer classes, we now clarify which officers may receive remedial relief. At one point, the district court's order suggests, with respect to the sergeant class, that relief be granted to all persons ranked higher on the promotional register than Gary Eugene, who received one of the sergeant promotions after being improperly certified under the "special certification" rule. Eugene ranked 107th on the register based on his exam score. There were over seventy officers who ranked ahead of Eugene on the register who were not selected for one of the sergeant promotions. To the extent the district court order is read in this way, it constitutes a clear abuse of discretion. None of the officers ranking higher than Eugene, who were not certified under the Rule of Eight, were eligible for the sergeant promotion under the City's civil service promotion rules. The City's unlawful "special certification," which resulted in the certification of Eugene, did not deprive any eligible sergeant candidates from being added to the qualified candidate pool. The City's discriminatory practice merely added additional minority candidates to the pool of eligible candidates as already determined by the written exam. The district court order attempts to fashion "make-whole" compensatory relief to those officers who were adversely affected by the City's discrimination. In the Title VII context, we have held unambiguously that "make-whole" relief is intended to recreate employment conditions that would have existed absent an employer's discrimination — i.e., to place "the injured party in the position he or she would have been in absent the discriminatory actions." Walters v. City of Atlanta, 803 F.2d 1135, 1145 (11th Cir. 1986) (citing Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975); Nord v. United States Steel Corp., 758 F.2d 1462, 1470 (11th Cir. 1985)). This holding draws support from the Supreme Court's conclusion in Stotts that Title VII "provide[s] make-whole relief only to those who have been actual victims of illegal discrimination." Firefighters Local Union No. 1784 v. Stotts, 467 U.S. 561, 580 (1984). The only police officers who were actually harmed by the City's discrimination were those officers who were "qualified" or eligible for the 1992 promotions according to the City's "Rule of Eight," and were passed-over in favor of the two ineligible officers. Twenty-three lieutenant and twelve sergeant officers fall into this group. Therefore, on remand, as we explain infra, we instruct that remedial relief only be granted to these thirty-five officers.
II.
We review the district court's remedial relief award pertaining to violations of the consent decree for an abuse of discretion. See Miranda v. B B Cash Grocery Store, 975 F.2d 1518, 1534 (11th Cir. 1992); EEOC v. Guardian Pools Inc., 828 F.2d 1507, 1511 (11th Cir. 1987).
District courts enjoy "wide discretion to fashion an equitable remedy for [civil] contempt that is appropriate to the circumstances." Guardian Pools, 828 F.2d at 1515. These sanctions may serve one of two broad purposes: (1) coercing the contemnor to comply with a court order, or (2) compensating a party for losses suffered as a result of the contemptuous act. See Jove Engineering Inc. v. IRS, 92 F.3d 1539, 1557 (11th Cir. 1996) (citing Guardian Pools, 828 F.2d at 1515). In serving these ends, a court's civil contempt power is measured solely by the "`requirements of full remedial relief.'" Citronelle-Mobile Gathering, Inc. v. Watkins, 943 F.2d 1297, 1304 (11th Cir. 1991) (quoting Guardian Pools, 828 F.2d at 1515) (citation omitted)). However, a district court may not use the civil contempt power to impose what amounts to a punitive or criminal contempt sanction. See In Re E.I. Dupont De Nemours Company-Benlate Litigation, 99 F.3d 363, 368 (11th Cir. 1996) (noting that it "`requires no citation of authority to say that a district court may not, even unwittingly, employ a civil contempt proceeding to impose what, in law, amounts to a criminal contempt sanction'") (quoting Blalock v. United States, 844 F.2d 1546, 1560 n. 20 (11th Cir. 1988) (per curiam) (Tjoflat, J., specially concurring)). A punitive or criminal contempt sanction may only be fashioned after many of the due process safeguards afforded to defendants in criminal proceedings — the right to counsel, the privilege against self-incrimination, the presumption of innocence, and the right to a jury trial in serious cases — are provided to an alleged contemnor. See Chandler v. James, 180 F.3d 1254, 1267 (11th Cir. 1999) (Tjoflat, J., specially concurring).
In this case, the district court conducted a civil rather than a criminal contempt proceeding. The district court ordered sweeping "compensatory" relief designed to "make-whole" those police officers adversely affected by the City's racially discriminatory certification procedures. Appellants do not challenge the district court's findings of racial discrimination and civil contempt. Therefore, the sole issue before us is whether the district court's broad remedial relief was an abuse of discretion. In answering this question, we draw on relevant principles of Title VII jurisprudence for evaluating "make-whole" compensatory relief.
The clear purpose of "make-whole" relief in the face of unlawful discrimination, whether in the Title VII or civil contempt context, is to "`recreate the conditions and relationships that would have been had there been no `unlawful discrimination.'" See International Bd. of Teamsters v. United States, 431 U.S. 324, 372 (1977) (quoting Franks v. Bowman Transp. Co., 424 U.S. 747, 769 (1976)). Regrettably, no ex post method of fashioning remedial relief can ever truly recreate the past as it would have existed absent an employer's discriminatory conduct. Nevertheless, in this case, we must attempt to discern how the 1992 police promotions likely would have proceeded in the absence of the discriminatory "special certification" procedures implemented by the City. In one respect, our task is clear. We know the qualified pool of eligible candidates, based on the promotional exam results, who were bypassed for promotion. All told, there were twenty-three certified lieutenant candidates and twelve certified sergeant candidates not promoted. We also know that one promotion for each of these ranks was received by a candidate who would not have been eligible absent the City's discrimination. Therefore, it is plain that, in the absence of City discrimination, only one of these twenty-three lieutenant candidates and only one of these twelve sergeant candidates would have received a promotion.
However, it is difficult, if not altogether impossible years later, to discern which of these candidates would have been promoted. The Miami Police Department utilized a wholly subjective interview process to make its final promotion decisions. A candidate's exam score played no role in his ultimate selection; it merely determined who would be "certified" or eligible for the final interview stage. Because of the subjective promotion process, the district court, quite apparently, could not identify which two officers, out of a total of thirty-five, should have received the 1992 promotions, and therefore it opted to employ a classwide promotion remedy instead. Previously, we have held that a classwide remedy may be employed in such circumstances. We have explained in the context of remedial backpay relief that a classwide remedy is appropriate when fashioning an individualized remedy would create a "`quagmire of hypothetical judgment[s]'" as to which individuals, out of a large class, should receive remedial relief. United States v. United States Steel Corp., 520 F.2d 1043, 1055 (5th Cir. 1975) (quoting Pettway v. American Cast Iron Pipe Co., 494 F.2d 211, 260 (5th Cir. 1974)). In endorsing this approach, we have recognized that the only other relief alternatives would be unpalatable: either (1) randomly selecting several individuals from a large class for full "make-whole" relief, or (2) awarding no relief at all because specific individuals deserving of a "make-whole" remedy could not be identified from a victim class. As a result, we have observed that remedial relief does not require "`unrealistic exactitude,'" and that "`uncertainties'" in the relief process "`should be resolved against the discriminating employer.'" United States Steel, 520 F.2d at 1050 (quoting Pettway, 494 F.2d at 260-61) (footnote omitted)). However, at the same time, we have instructed that classwide remedies must strive for equity to both parties, and that the "key is to avoid both granting a windfall to the class at the employer's expense and the unfair exclusion of claimants by defining the class . . . too narrowly." Pettway, 494 F.2d at 262 n. 152.
Decisions by the former Fifth Circuit prior to September 30, 1981 constitute binding precedent in our circuit. See Bonner v. Prichard, 661 F.2d 1206, 1207 (11th Cir. 1981).
While we agree with the classwide approach employed by the district court under the facts of this case, we find that the scope of the remedy imposed plainly was overly broad. Simply put, the district court treated each bypassed candidate as if he had a one hundred percent probability of receiving a promotion absent the City's discrimination. Each eligible promotion candidate received the full value of a 1992 promotion including backpay, retroactive seniority, a fifteen thousand dollar pension payment, and a one rank promotion from their current position. However, the district court itself acknowledged that each candidate stood much less than a one hundred percent chance of promotion since it could not identify two individuals who would have been likely promoted absent the City's discrimination. Based on mathematical probability alone, each lieutenant candidate stood only a one in twenty-three (or four percent) chance of promotion, and each sergeant candidate stood only a one in twelve (or eight percent) chance of promotion.
Given these very long odds, we conclude that the district court should have relied on a variant of the pro rata method for computing classwide remedial relief. This computation method awards each class member a proportional share of the full monetary value of the promotion for which they were eligible. In this case, each certified sergeant candidate not selected for the promotion would receive a one-twelfth share; while each certified lieutenant candidate not selected for the promotion would receive a one-twenty-third share. In the past, we have recommended this specific method when, even after reasonable effort, a court is unable to differentiate between class members for remedial relief purposes. See United States Steel, 520 F.2d at 1055-56. In United States Steel, the former Fifth Circuit "commended" to the district court a pro rata methodology where there was clear proof of systematic race discrimination against a large class of black steelworkers but an unclear methodology for determining which specific individuals deserved backpay relief. See United States Steel, 520 F.2d at 1055-56. Several of our sister circuits also have recommended the pro rata method in similar circumstances. See Dougherty v. Barry, 869 F.2d 605, 614-15 (D.C. Cir. 1989) (directing district court to employ pro rata method where eight plaintiffs would have competed, absent defendant's discrimination, for two promotions); Ingram v. Madison Square Garden Center, Inc., 709 F.2d 807, 812 (2d. Cir. 1983) (vacating individual "make-whole" relief to a victim class that exceeded the number of actual vacancies and recommending pro rata method instead); Hameed v. International Ass'n of Bridge, Structural, Ornamental Iron Workers, Local Union No. 296, 637 F.2d 506, 519-21 (8th Cir. 1980) (recommending classwide pro rata backpay award given difficulty of determining which individuals, out of 45 plaintiffs, would have received apprenticeships absent discrimination); Stewart v. General Motors Corp., 542 F.2d 445, 452-54 n. 7 (7th Cir. 1976) (favoring classwide pro rata award where defendant's subjective promotion procedures made it impossible to fashion an individualized remedy). We find the factual posture of Dougherty particularly germane. In Dougherty, eight white firefighters received promotions as compensatory relief for the District of Columbia Fire Department's racially discriminatory promotion practices, even though only two of the firefighters would have received a promotion absent the discriminatory practices. The court of appeals vacated the award, since each class member stood only a one in four chance of promotion, and instead recommended that each firefighter receive a pro rata share of the promotions' value. Dougherty, 869 F.2d at 614 (concluding that "in order to restore appellees to the position they would have occupied absent discrimination, the district court should have awarded each appellee a fraction of the promotions' commensurate value with the likelihood of his receiving one of the promotions") (citation omitted).
The only notable exception to this approach occurred in Taxman v. Board of Educ., 91 F.3d 1547 (3d Cir. 1996), where the Third Circuit upheld full backpay relief to a teacher laid off as a result of a discriminatory school board affirmative action policy — even though the plaintiff would have maintained only a fifty-fifty chance of retaining her job absent the discrimination. See id. at 1565-66. In deciding between two teachers of equal seniority and merit for one position, the school board used race to break the tie, rather than a coin toss. The court of appeals concluded that because the plaintiff had a fifty percent probability of retaining her position and the school had unclean hands as the discriminatory wrongdoer, it was within the district court's discretion to award full backpay. See id. This case is distinguishable since it involved two individuals for one position, and a probability of retention (fifty percent) closely approximate to a more likely than not standard, rather than a very large class vying for only one position.
In this case, each class member stood a much slimmer chance of obtaining one of the two available promotions. Only one of twelve sergeant candidates (roughly eight percent of the sergeant class) and one of twenty-three lieutenant candidates (roughly four percent of the lieutenant class) actually would have received a promotion absent the City's discrimination. In short, the district court's award placed the officers in the class in a substantially better position then they would have occupied absent the City's discrimination by dramatically inflating their actual chances of obtaining only one of two available positions. While cognizant of the need to compensate victims of discrimination, even where relief calculations are complicated by the size of the class, we conclude that "make-whole" remedial relief awards, at the very least, must be proportionate to a court's best determination of the actual compensatory losses of a class. In this way, we balance between our twin duties to fairly compensate discrimination victims and yet avoid punitive remedial relief awards.
Under these facts, full remedial relief for all thirty-five officers — which is what the district court ordered — amounts to an unfair and sweeping windfall to the officer class and an abuse of the district court's discretion to award remedial relief. Its estimated cost would be around nine million dollars. However, the estimated monetary value of the two lost promotions is only a little over five hundred thousand dollars. To be sure, the fashioning of remedial relief is a highly fact-specific inquiry best left to the discretion of district courts. That said, the unambiguous goal of "make-whole" relief is to compensate persons who have actually suffered from unlawful discrimination for their actual compensatory losses. Therefore, remedial relief must be constructed with a close eye towards an employer's likely promotion practices and those class members likely to have been promoted had there been no discrimination. See Ingram, 709 F.2d at 812 (observing that "remedial relief should be granted only to those class members who would have filled vacancies had there been no discrimination"). We find the nexus between thirty-five candidates and only two promotion slots far too disproportionate to sustain full remedial relief for the entire class.
The estimated backpay differential is $68,000 for each lieutenant and $58,000 for each sergeant, totaling around $2.2 million dollars. The lump-sum pension payments amount to $525,000 ($15,000 for each of the thirty-five officers). The City Retirement Trust also estimates that lost retirement benefits, for the entirety of an officer's career, would amount to another $180,000 for each officer or $6.3 million dollars for the entire officer class.
In fact, the amount awarded by the district court so exceeds the monetary value of the lost promotions that we can only interpret the award as punitive in nature. The "compensatory" civil contempt class award is eighteen times greater than the actual value of the two lost promotions. While we previously have recognized that the line between civil and criminal contempt sanctions is not "always clear," Chandler, 180 F.3d at 1267 (Tjolfat, J., specially concurring), and that we must draw such conclusions from "`the character of the relief itself,'" International Union, United Mine Workers v. Bagwell, 512 U.S. 821, 828 (1994) (citation omitted), this "compensatory" award plainly is so excessive as to be punitive. Moreover, the district court's award of thirty-five retroactive promotions, where absent the City's discrimination only two additional promotions would have been available, could radically restructure the City's police force by creating many more lieutenants and sergeants than the City sought fit to create under its own promotion policies. The very magnitude of this remedy risks reshaping the Police Department in a variety of ways unforseen and unintended by the district court.
We therefore vacate the award, and remand the matter to the district court. Because of the subjective nature of the City's promotion process, the district court was unable to identify two individual class members who should have received the 1992 promotions. On remand, we therefore direct the district court to award each certified officer candidate a pro rata share of the monetary value of the promotion for which they were eligible.
At oral argument, both parties conceded that the City of Firefighters' and Police Officers' Retirement Trust, who played no role in the City's discriminatory practices and who administer retirement benefits to all former City of Miami police officers, should not be responsible for the monetary value of the officers' lost pension benefits. To the extent the district court awards the thirty-five officers a pro rata share of lost pension benefits, the City, and not the Trust, shall be held liable for those monies.
III.
In sum, the district court's remedial award was excessive, and should have been limited to a pro rata division among those officers certified for the 1992 promotions. The twelve sergeant candidates should share, on a pro rata basis, the value of the sergeant promotion. The twenty-three lieutenant candidates should share, on a pro rata basis, the value of the lieutenant promotion. Accordingly, we vacate the district court's award and remand.
VACATED IN PART AND REMANDED.