Dean v. City of Shreveport

United States Court of Appeals Fifth Circuit F I L E D REVISED FEBRUARY 10, 2006 UNITED STATES COURT OF APPEALS January 25, 2006 FOR THE FIFTH CIRCUIT Charles R. Fulbruge III Clerk

No. 04-31163

JEFFREY TODD DEAN, ET AL.

Plaintiffs-Appellants,

VERSUS

THE CITY OF SHREVEPORT,

Defendant-Appellee.

Appeal from the United States District Court For the Western District of Louisiana ( 5:00-CV-2372 )

Before GARWOOD, SMITH, and DeMOSS, Circuit Judges. DeMOSS, Circuit Judge:

INTRODUCTION

Plaintiffs-Appellants (“Appellants”) challenge the district court’s grant of summary judgment in favor of Defendant-Appellee City of Shreveport (the “City”) dismissing Appellants’ 42 U.S.C. § 1983, Title VII, and Louisiana constitutional and statutory claims. We affirm in part, reverse in part, and remand the case for further proceedings consistent with this opinion.

Appellants are white males who were denied employment after applying to become City firefighters. At the time Appellants *2 applied, the City used a hiring process that placed applicants into separate lists according to race and sex. The City created its race-conscious hiring process in an attempt to comply with a 1980 consent decree drafted to end discriminatory hiring practices in the City’s fire department and to remedy the effects of past discrimination. Appellants challenge both the decree and the hiring process.

FACTUAL BACKGROUND

In 1977, the U.S. Department of Justice (“DOJ”) filed a lawsuit against the City alleging its fire department used racially and sexually discriminatory hiring practices. To settle the lawsuit, the City signed a proposed consent decree providing a plan to end then-current discriminatory practices and remedy the effects of past discrimination. Although the City signed the decree, it did not admit to any unlawful discrimination. Because the City declined to admit to unlawful discrimination, the district court initially refused to enter the decree. United States v. City of Alexandria, No. 77-2040, 1977 WL 69 (E.D. La. July 22, 1977). However, in 1980, this Court reversed the district court and ordered the decree be entered. United States v. City of Alexandria, 614 F.2d 1358 (5th Cir. 1980).1

1 We reviewed the decree at that time under a rat ional basis standard of review. City of Alexandria, 614 F.2d at 1363 (inquiring whether the decree was “reasonably related to the legitimate state goal of achieving equality of employment opportunity”). This standard of review no longer applies, and we now strictly scrutinize all race-conscious remedies to ensure they are narrowly tailored to achieve a compelling government interest. City of Richmond v. J.A. Croson Co., 488 U.S.

To remedy the effects of past discrimination, the decree sets forth a long-term goal that the City achieve – subject to the availability of qualified applicants – the same proportions of blacks and women in its fire department “as blacks and women bear to the appropriate work force in the particular jurisdiction.”2 However, the decree does not define “appropriate work force.” The decree also requires the City to adopt an interim hiring goal of filling at least fifty percent of all firefighter vacancies with qualified black applicants and at least fifteen percent with qualified female applicants. The interim goal remains in effect until the long-term goal is achieved and maintained for one year.

The decree itself does not mandate any particular hiring process for meeting its goals. Therefore, the City formed its own process.3 Phase one requires all firefighter applicants to take the Civil Service Exam. To pass, an applicant needs a score of at

469, 493-94 (1989). Thus, as we re-evaluate the decree under strict scrutiny, we are not bound by our prior approval of it under the rational basis standard. 2 The decree is published as an appendix to City of Alexandria, 614 F.2d at 1367-72. 3 The City’s hiring process remained substantially the same from the time the decree was entered until Appellants were denied employment between 2000 and 2002. In 2004, the City changed its hiring process. The City claims it continues to strive toward the interim and long-term goals in the decree, but that its new hiring process is race-neutral. Much to this Court’s dissatisfaction, the City has kept secret the details of its new hiring process. In addition, the City has failed to explain how it expects its new allegedly race-neutral hiring process to meet the goals of the decree when its old race-conscious process apparently did not. In any event, we limit our analysis to the hiring process the City used at the time Appellants were denied employment because it is that process Appellants argue violated their rights. We refer to the City’s old hiring process in the present tense in our opinion purely to avoid continuous, and possibly confusing, tense changes.

least seventy-five. Points are then added to the scores of applicants with prior emergency medical or paramedic training or military service. When the final numerical scores are calculated, the applicants are separated into three lists: a white male list, a black male list, and a female list. Each list is ranked by exam score from highest to lowest. The City then determines how many firefighter positions it needs to fill. Finally, starting with the highest exam score on each list, the City selects approximately twice as many applicants as vacant spots to proceed to phase two of the hiring process. Of those selected to proceed, fifty percent of the males are white and fifty percent are black. Every female who receives a seventy-five on the exam usually proceeds to phase two because of the extremely low number of female applicants.

Phase two includes six additional steps an applicant must pass to become a firefighter: (1) an agility test; (2) a general preliminary interview, screening for disqualifying conduct, such as drug use; (3) a criminal background check; (4) a polygraph exam; (5) a psychological exam and interview; and (6) a medical exam. An applicant who fails any step is immediately denied employment. An applicant who passes each step is immediately awarded employment.

Under this hiring process, Appellants’ exam scores in phase one were too low on the white male list to proceed to phase two.

PROCEDURAL HISTORY

In October 2000, Appellant Jeffery Todd Dean (“Dean”), an *5 unsuccessful white male applicant, sued the City under 42 U.S.C. § 1983, alleging the decree and the hiring process violate the Equal Protection Clause of the Fourteenth Amendment. Dean and the City filed cross-motions for summary judgment. In 2002, during the pendency of those motions, eight additional white male applicants brought similar discrimination suits against the City. In addition to Dean’s equal protection claim, they asserted claims under Title VII, the Louisiana Constitution, and a Louisiana anti-discrimination employment statute.4 The cases were consolidated with Dean’s, and all eight joined his pending motion for summary judgment. All parties stipulated to proceed before a magistrate judge. In October 2004, the magistrate judge denied Dean’s motion for summary judgment and granted the City’s, dismissing all claims against the City. This timely appeal followed.

DISCUSSION

I. Applicable Standard of Review

We review a district court’s grant of summary judgment de novo and apply the same standards as the district court. Daniels v. City of Arlington, 246 F.3d 500, 502 (5th Cir. 2001). Summary judgment is proper only “if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any

4 We address a total of four claims on appeal. However, not all nine Appellants brought each of these claims below. Therefore, on remand the district court’s first order of business will be to determine the effects of our decision today on each individual Appellant.

material fact and that the moving party is entitled to judgment as a matter of law.” FED. R. CIV. P. 56(c); see also Priester v. Lowndes County, 354 F.3d 414, 419 (5th Cir. 2004). “If the moving party meets the initial burden of showing there is no genuine issue of material fact, the burden shifts to the nonmoving party to produce evidence or designate specific facts showing the existence of a genuine issue for trial.” Priester, 354 F.3d at 419. We resolve doubts in favor of the nonmoving party and make all reasonable inferences in favor of that party. Id. II. Appellants’ Equal Protection Clause Claim

Appellants first contend that the consent decree and the City’s hiring process violate their right to equal protection under the United States Constitution.

a. Strict Scrutiny

The Equal Protection Clause of the Fourteenth Amendment provides that “[n]o State shall . . . deny to any person within its jurisdiction the equal protection of the laws.” U.S. CONST. amend. XIV, § 1 (emphasis added); see also City of Richmond v. J.A. Croson Co., 488 U.S. 469, 493 (1989). “Classifications based on race carry a danger of stigmatic harm . . . [and] may in fact promote notions of racial inferiority and lead to a politics of racial hostility.” Croson, 488 U.S. at 493. Thus, all race-conscious measures receive strict scrutiny review under the Equal Protection Clause. See id. at 493-94; see also Black Fire Fighters Ass’n v. *7 City of Dallas, 19 F.3d 992, 995 n.6 (5th Cir. 1994) (explaining that this standard applies to consent decrees). Strict scrutiny review demands that a race-conscious measure be (1) justified by a compelling government interest and (2) narrowly tailored to further that interest. Police Ass’n ex rel. Cannatella v. City of New Orleans, 100 F.3d 1159, 1167 (5th Cir. 1996).

1. Compelling Interest

It is well settled that the government has a compelling interest in remedying its own past discrimination. See United States v. Paradise, 480 U.S. 149, 167 (1987). However, a general assertion of past societal discrimination is insufficient. See Croson, 488 U.S. at 499. Rather, the government must justify its action with a showing of past discrimination by the governmental unit seeking to use the race-conscious remedy. See id. at 495-97; see also Wygant v. Jackson Bd. of Educ., 476 U.S. 267, 276 (1986); Police Ass’n, 100 F.3d at 1168.

The Supreme Court has offered little guidance as to how much evidence of past discrimination is required. Id. However, “[t]here is no doubt that ‘[w]here gross statistical disparities can be shown, they alone in a proper case may constitute prima facie proof of a pattern or practice of discrimination.’” Croson, 488 U.S. at 501 (alteration in original) (quoting Hazelwood Sch. Dist. v. United States, 433 U.S. 299, 307-08 (1977)). The relevant statistical comparison is between the number of minorities in the *8 work force of the governmental unit and “the number of minorities qualified to undertake the particular task.” See Croson, 488 U.S. at 502.

Appellants argue that even in 1980, the City had no compelling interest to justify the decree or a race-conscious hiring process. Appellants point out that the decree is not based on any formal factual finding of past discrimination. Further, the City’s hiring process was adopted solely to comply with the goals of the decree. Appellants urge us to hold that a governmental unit may use a race-conscious remedy only after a formal judicial, legislative, or administrative finding of past discrimination. The City maintains that in 1980 it clearly had a compelling interest. The City concedes that a formal finding of past discrimination was never made, but insists that one was not required.

We agree with the City. Nothing in Supreme Court or Fifth Circuit precedent compels us to require a formal finding of discrimination prior to the use of a race-conscious remedy. To the contrary, in Police Ass’n, 100 F.3d at 1167-68, we stated that “Croson does not require a city to incriminate itself by proving its own participation in past discrimination.” It is when a remedial program is challenged that a trial court must make a factual determination that there was a strong basis in evidence for the conclusion that remedial action was necessary. Thus, what we meant in Police Ass’n was that the government need not incriminate *9 itself with a formal finding of discrimination prior to using a race-conscious remedy, but if the remedy is later challenged, a court must determine there was a strong evidentiary basis for its enactment. This approach is consistent with pre-Croson Supreme Court precedent. See Wygant, 476 U.S. at 277 (explaining that “[e]videntiary support for the conclusion that remedial action is warranted becomes crucial when the remedial program is challenged in court by nonminorit[ies]” (emphasis added)); see also id. at 289 (O’Connor, J., concurring) (agreeing that “a contemporaneous or antecedent finding of past discrimination . . . is not a constitutional prerequisite . . . to an affirmative action plan” (emphasis added)). This approach is also in line with other circuits that have addressed the issue. See, e.g., Aiken v. City of Memphis, 37 F.3d 1155, 1162-63 (6th Cir. 1994) (“No formal finding of past discrimination by the government unit involved is necessary . . . .”); see also In re Birmingham Reverse Discrimination Employment Litig., 20 F.3d 1525, 1539 (11th Cir. 1994) (explaining that the government “was not required to make formal findings about its own past discrimination--it merely had to have a strong basis in evidence”(emphasis added)). Thus, to the extent our prior decisions were unclear, we now clarify that when a governmental unit employs a race-conscious remedy, it need not have already made a formal finding of past discrimination. Nevertheless, if the remedy is later challenged, the reviewing *10 court must ensure there was strong evidence of past discrimination warranting the remedy. We turn now to whether strong evidence of discrimination existed in this case, such that in 1980 the City had a compelling interest in remedying it.

The district court concluded that the City made an adequate showing of past discrimination. We agree. Prior to 1974, the City’s fire department had never hired a black employee. In 1974, after it was sued by black applicants alleging racially discriminatory hiring practices, the City hired three black firefighters. After the lawsuit was settled, the City hired no black employees in 1975, just one black firefighter in 1976, and no black employees in 1977. In 1977, another lawsuit was brought against the City, this time by the DOJ, alleging racially and sexually discriminatory hiring practices. In the time between the 1977 lawsuit and the 1980 decree, the City hired only six additional black firefighters. Thus, when the decree was entered to settle the DOJ lawsuit, only 10 of the City’s 270 firefighters were black.5 At that time, blacks accounted for approximately forty percent of the general population in the City of Shreveport. Further, the City now admits it systematically excluded all black applicants prior to 1974 and hired the few it did between 1974 and 1980 in response to pending lawsuits. Appellants do not contest these statistics and offer no alternative explanation for them,

5 By 1980, the City still had not hired a single female firefighter.

expert or otherwise.

We recognize that the relevant comparison when determining whether discrimination existed is between the number of black firefighters in the City’s fire department and the “number of [blacks] qualified to undertake the particular task,” see Croson, 488 U.S. at 502, not the number of blacks in the general population, see id. We are also aware that the City has not presented a precise calculation of how many black applicants were qualified to become firefighters in the years leading up to the 1980 decree. Nevertheless, we find it inconceivable that the number of qualified blacks was an “emphatic zero.” See Dean v. City of Shreveport, No. 00-2372, slip op. at 14 (W.D. La. Oct. 22, 2004) (quoting Guice v. Fortenberry, 661 F.2d 496, 505 (5th Cir. 1981) (en banc)).

In most cases, a governmental unit’s failure to provide statistical data comparing the number of minorities in its work force with the number of minorities qualified to undertake the particular task, rather than the number of minorities in the general population, will prove fatal to an attempt to show past discrimination. See, e.g., Croson, 488 U.S. at 501-02 (finding no compelling interest in remedying past discrimination because the government failed to show how many minorities in the relevant market were qualified to undertake the particular task). But in rare cases, the statistical disparity may be so great between a *12 particular work force and the general population that, along with other overwhelming evidence, it may provide us with an adequate basis to conclude no genuine factual issue remains regarding the existence of past discrimination.

This is such a case. In addition to the fact that the City hired no black employees prior to 1974 and only 10 black firefighters as of 1980, the overwhelming evidence shows that (1) the City now admits that for over 100 years it systematically excluded all minorities from its fire department; (2) the City has been sued numerous times for racial and sexual discrimination; and (3) Appellants have failed to offer any alternative explanation, expert or otherwise, for the gross statistical disparity. Therefore, the district court properly concluded that the City had a compelling interest in 1980 to enter into the decree and implement a race-conscious hiring process.6

The City argues that this conclusion ends our compelling interest inquiry. We disagree. In addition to showing past

6 Again, we do not intend to suggest a statistical disparity between a work force and the general population, without additional overwhelming evidence, is enough to show past discrimination. Had the City not admitted to systematically excluding blacks from its work force and twice previously been sued for racial discrimination, our decision today might have been different. See Paradise, 480 U.S. at 169 (taking into account that the governmental unit systematically excluded minorities and had previously faced numerous allegations of racial discrimination). Similarly, our decision might have been different had Appellants rebutted the City’s evidence or offered an alternative explanation, thereby raising some genuine factual issue. See Priester v. Lowndes County, 354 F.3d 414, 419 (5th Cir. 2004) (explaining that after the moving party meets it initial burden of showing no genuine issue of material fact remains, the burden shifts to the nonmoving party to raise one by producing evidence or by pointing to specific facts).

discrimination, the City must also convince us that when Appellants were denied employment between 2000 and 2002, lingering effects of past discrimination still necessitated a race-conscious remedy. See Paradise, 480 U.S. at 169-70 (focusing repeatedly on whether any remaining effects of past discrimination still warranted a race-conscious remedy); see also Police Ass’n, 100 F.3d 1168-69 (same). If the effects of past discrimination no longer existed when Appellants were denied employment, the City no longer had a compelling interest to justify a race-conscious remedy. See Paradise, 480 U.S. at 169-70; see also Police Ass’n, 100 F.3d 1168-69. The district court in error focused solely on whether the City had a compelling interest in 1980. See Dean, slip op. at 13-15. Thus, we must determine on appeal whether any effects of the City’s past discrimination still existed when Appellants were denied employment, such that the City still had a compelling interest to justify its race-conscious remedy at that time.

Again, as mandated by Croson, 488 U.S. at 501-02, the relevant statistical comparison is between the number of blacks in the City’s fire department and the number of blacks qualified to undertake the particular task. Thus, in order to demonstrate that its remedy was still necessary between 2000 and 2002, the City had to show what percentage of its qualified applicant pool was black during that time period. The City did not do so. The record is entirely inconsistent regarding this issue, and the district court *14 failed to address it.

For example, Fire Chief Kelvin J. Cochran (“Chief Cochran”) was asked in his deposition, “Do you know of any study or anything that would indicate what the pool of available qualified applicants might be blacks and women?” He responded, “No, ma’am.” Chief Cochran was later asked, “To your knowledge, is there any kind of study available . . . that would give some information on what the actual demographics are for your fire district area or your hiring area?” He again responded, “No, ma’am.”

Attempting to provide this crucial data, the City later hired an expert statistician to calculate the demographics of its qualified labor pool.7 The expert’s conclusions rested on the assumption that the percentage of blacks passing the Civil Service Exam established the percentage of blacks in the qualified labor pool. We see numerous problems with this assumption, all of which will require careful consideration on remand.

First, it is inappropriate to rely on an expert statistician with a Ph.D. in Economics to determine what makes an applicant qualified to become a firefighter. The City, its fire department, or a vocational expert must make this determination. See Olson v. Schweiker, 663 F.2d 593, 596-97 (5th Cir. 1981) (using a vocational

7 The DOJ also attempted to determine whether a statistical disparity existed between the fire department’s work force and its qualified labor pool. After an initial conclusion in 2002 that there was no disparity, the DOJ later concluded it “lacked [the] essential data necessary to render a determination.” This illustrates the need for further factual development on remand.

expert to determine whether plaintiff was qualified for particular positions); see also New Orleans (Gulfwide) Stevedores v. Turner, 661 F.2d 1031, 1035 (5th Cir. 1981) (same); Simons v. Sullivan, 915 F.2d 1223, 1224 (8th Cir. 1990) (same); Noble v. Ala. Dep’t of Envtl. Mgmt., 872 F.2d 361, 363 (11th Cir. 1989) (deferring to the state agency when establishing what qualifications were required for certain state positions). A statistician, after he is informed what a qualified applicant is, may then calculate the demographics of the qualified labor pool. Second, the fire department itself contradicts its own expert’s definition of a qualified applicant. The expert defined qualified applicants as those having passed the Civil Service Exam. Yet the City maintains, and the district court stated in its ruling, “there is no evidence that persons with higher passing scores make better fire[fighters] than those with low passing scores.” Dean, slip op. at 17. We fail to understand how passing scores conclusively establish the demographics of the qualified applicant pool if passing scores mean nothing with respect to predicting the quality of future firefighters. Third, logic cuts against equating the percentage of blacks who pass the Civil Service Exam with the percentage of blacks in the qualified labor pool. Equating the two percentages ignores the six subsequent steps an applicant must successfully pass in order to become a firefighter, including: (1) an agility test; (2) a general preliminary interview, screening for disqualifying conduct, such as *16 drug use; (3) a criminal background check; (4) a polygraph exam; (5) a psychological exam and interview; and (6) a medical exam. Surely, an applicant who receives a passing score but then fails one of these requirements is not qualified.

On remand, the City must properly define a “qualified applicant.” It must then provide reliable statistical data showing the percentages of blacks in its work force and in its qualified labor pool between 2000 and 2002. Only when the district court has this information can it properly decide whether a sufficient disparity still existed. Until then, a genuine issue of material fact remains, thereby preventing a legal analysis of whether the City’s race-conscious remedy was still necessary between 2000 and 2002. Although this alone requires us to reverse the district court’s judgment and remand the case for further factual development, we will also review the court’s narrow tailoring analysis.

2. Narrow Tailoring

The Supreme Court has focused on the following factors (“Paradise factors”) when reviewing a race-conscious remedy to ensure it is narrowly tailored: the necessity of the particular relief and the efficacy of alternative remedies; the flexibility and duration of the relief, including the availability of waiver provisions; the relationship between the numerical goal of the relief and the relevant labor market; and the impact of the relief *17 on the rights of third parties. See Paradise, 480 U.S. at 171; see also Black Fire Fighters Ass’n, 19 F.3d at 995. Considering these factors, it is clear that additional factual issues still remain, such that we cannot properly conduct a narrow tailoring analysis on this record. However, we briefly discuss each Paradise factor to highlight legal errors in the district court’s analysis and specific factual disputes that must be resolved on remand.

a. Necessity of the Particular Relief and Efficacy of Alternative Remedies

The district court concluded that the decree’s interim hiring goal and the City’s hiring process were necessary remedies. The court also decided that alternative remedies would have been insufficient to remedy past discrimination. For the same reason that we cannot complete our compelling government interest analysis at this time--namely, the City’s failure to explain the demographics of its qualified labor pool--we cannot properly review and weigh the first Paradise factor. We will not be able to determine whether the remedies in this case were necessary between 2000 and 2002, or whether alternative remedies would have sufficed, until we know what statistical disparity still existed at that time, if one still existed at all, between the fire department’s work force and its qualified labor pool.

However, before moving on to the next Paradise factor, we must point out numerous legal errors in the district court’s analysis of this factor. The court reasoned that “[t]he absolute lack of black *18 or female fire[fighters] prior to the institution of [the 1973] litigation, followed by a lapse in minority hiring until . . . the [1980] consent decree,” made strong remedial measures necessary. See Dean, slip op. at 16. This may have been true in 1980. However, these events alone do not illustrate whether strong measures were still necessary when Appellants were denied employment. To the extent the district court focused on whether the remedies were necessary in 1980, instead of between 2000 and 2002, it erred. There is no legal basis for the conclusion that because a particular race-conscious remedy was necessary at one point in time, it is still necessary two decades later. See Paradise, 480 U.S. at 171-72 (focusing on whether the race-conscious remedy was still necessary when plaintiffs were denied promotions); see also Police Ass’n, 100 F.3d 1169 (same). On remand, the City must show that the decree and hiring process were necessary when Appellants were denied employment between 2000 and 2002.

The district court similarly erred in its analysis of the efficacy of alternative remedies. By focusing on the time leading up to the 1980 decree, the court could not have properly analyzed whether alternative remedies would have sufficed when Appellants were denied employment. The fact that alternative measures would have been insufficient in 1980 does not indicate whether they would have been insufficient when Appellants were denied employment. On *19 remand, the City must show that race-neutral or less intrusive remedies would have been insufficient between 2000 and 2002.

In addition, the record is currently too inconsistent to determine what alternative remedies, if any, the City has already attempted and whether those or any others will suffice. For example, one alternative to race-conscious hiring is increased recruiting efforts targeting minorities. In his deposition, Chief Cochran stated that when the City recruits, it “does not specifically target black and female applicants.” But in the same deposition he stated that the City specifically targets blacks by “targeting the African-American churches.” When asked whether the City’s recruiting policy is adequately geared towards attracting minorities, Chief Cochran admitted, “the City has never done any kind of self-evaluation to see if its recruiting efforts are appropriate for recruiting minorities and women.” Thus, the efficacy of alternative measures remains a genuine issue of material fact that must be resolved on remand.

b. Flexibility and Duration

The district court found the decree and the hiring process adequately flexible. The court also decided that their long duration did not preclude finding the remedies narrowly tailored.

The primary question when analyzing a remedy’s flexibility is whether its requirements may be waived. See Paradise, 480 U.S. at 177 (focusing its flexibility analysis solely on whether *20 requirements could be waived). If they may, the remedy is adequately flexible. See id. The decree requires the City to adopt the long-term goal of achieving, subject to the availability of qualified applicants, the same proportions of blacks and women in its fire department “as blacks and women bear to the appropriate work force in the particular jurisdiction.” It also requires the City to adopt an interim goal of hiring at least fifty percent black and fifteen percent female employees until the long-term goal is achieved and maintained for one year. The goal of the hiring process parallels the interim goal in the decree. While the decree does not allow the goals to be waived, it does specify that they are “subject to the availability of qualified applicants.”

Despite this explicit exception, Appellants argue that the decree and hiring policy require a rigid fifty percent racial quota. The City, relying on the waiver provision, insists both are flexible. The City also points out that since 1980, it has hired less than fifty percent black employees in all but two hiring classes. We agree with the district court that the remedies are adequately flexible. We do so because the remedies here, as far as their flexibility is concerned, parallel the flexibility of the remedy in Paradise. In that case, the Alabama Department of Public Safety was required to award half of all state trooper promotions to black employees. Paradise, 480 U.S. at 153. However, the requirement was contingent on the availability of qualified *21 candidates. Id. at 177. This alone satisfied the flexibility requirement. Id. at 177-78. Like the remedy in Paradise, the remedies here are contingent on the availability of qualified applicants and are therefore adequately flexible.

The central theme of a duration analysis is that the shorter the life-span of the remedy, the more likely it is narrowly tailored. See Paradise, 480 U.S. at 178. The City’s obligations under the decree end when it achieves its long-term goal, that is, when the City achieves the same proportions of blacks and women in its fire department “as blacks and women bear to the appropriate work force.” In order to estimate when a particular goal might be achieved, the goal must be clear. Thus, here it must be clear what proportions “blacks and women bear to the appropriate work force.” Determining this requires a precise definition of the phrase “appropriate work force.” We have no precise definition.8

8 Contrary to Appellants’ contention, we did not define the “appropriate work force” in United States v. City of Alexandria, 614 F.2d 1358 (5th Cir. 1980). However, we did state that the decree “established long term goals of achieving . . . the same percentages of blacks and women as are present in the workforces in the various affected localities.” Id. at 1362. Appellants argue this statement conclusively established the goal of the decree as achieving a percentage of blacks in the fire department equal to that in the general population. Appellants also insist we are now bound by that statement. This would make the goal of the decree unconstitutional under Croson, 488 U.S. at 501-02. We disagree with Appellants for several reasons. First, by elevating the standard of review to strict scrutiny, Croson, 488 U.S. at 494, effectively overruled our holding in City of Alexandria. Second, the definition of the phrase “appropriate work force” was not at issue in that case. See generally City of Alexandria, 614 F.2d at 1361-67. Third, our statement referring to “the workforces,” id. at 1362, may have been shorthand for “the appropriate work forces,” to which the plain language of the decree refers. In any event, we are not bound today by our alleged interpretation of the decree, or our subsequent approval of it, in City of Alexandria.

Therefore, the City’s goal remains unclear, and we cannot properly analyze duration.

When interpreting a consent decree, general principles of contract interpretation govern. United States v. Chromalloy Am. Corp., 158 F.3d 345, 349 (5th Cir. 1998). We begin by looking to the “four corners” of the decree. Id. at 350. We will then look to extrinsic evidence if the decree is ambiguous. N. Shore Lab. Corp. v. Cohen, 721 F.2d 514, 519 (5th Cir. 1983). A decree “is ambiguous when it is reasonably susceptible to more than one meaning, in light of surrounding circumstances and established rules of construction.” Id.

We find the phrase “appropriate work force” ambiguous as a matter of law. The decree itself sheds no light on what the adjective “appropriate” means in this context. Because we cannot decipher its meaning within the four corners of the decree, we need extrinsic evidence to aid our interpretation. We have none. For example, we have no preliminary drafts of the consent decree or correspondence between the parties during negotiations. This evidence would help us determine what the parties meant by “appropriate work force.” Without extrinsic evidence, we have no reason to choose one possible definition over another. Because the phrase “appropriate work force” was not defined by the district court, and because we are unable to do so without extrinsic evidence, the issue must be resolved on remand.

We also stress that regardless of the outcome of the issue on remand, the duration component of this Paradise factor will favor Appellants, at least to some degree. The durations of the remedies in this case are breathtakingly long in comparison to others we have reviewed.9 Edwards v. City of Houston, 78 F.3d 983, 1002 (5th Cir. 1996) (en banc), involved a consent decree that allowed a police department to promote a certain number of minority officers to sergeant and lieutenant.10 The remedy was to last no longer than five years. Id. In Black Fire Fighters Ass’n, 19 F.3d at 997, we struck down a consent decree that allowed a fire department to promote a certain number of minorities to higher ranking positions. We found the remedy not to be narrowly tailored even though it lasted for only three years. Id. Finally, in Police Ass’n, 100 F.3d at 1173, we again struck down a race-conscious promotional plan. We found the remedy not to be narrowly tailored even though it was a one-time set of promotions, not an ongoing plan. Id. at 1169.

Therefore, the durations of the remedies in this case are already considerably longer than those in any of our previous cases. Thus, this factor will weigh in Appellants’ favor

9 In the midst of this litigation in 2004, the City changed its hiring process. Nevertheless, we cannot ignore the fact that when Appellants’ claims arose, the process had been used for over twenty years. The decree had been in effect for just as long, and continues to this day. 10 Although we did not reach the merits regarding the consent decree’s constitutionality in Edwards, we refer to the case to illustrate just how long the durations of the remedies in this case are in relation to others we have been confronted with in the past.

regardless of how soon the district court determines the remedies might end. We point this out only to emphasize that in order for these remedies to be upheld on remand, other Paradise factors must strongly suggest they are narrowly tailored.

On remand, the district court must determine, using extrinsic evidence, the decree’s precise long-term goal by defining the “appropriate work force.” It must then estimate how long it will take for the City to achieve that goal. When an approximate end date is known, the district court can then re-weigh the duration factor, keeping in mind the past precedent we have just highlighted.

c. Relationship Between the Numerical Goal and the Relevant Labor Market

The district court conducted no meaningful analysis of this factor. Nor could it have. The long-term numerical goal of a race-conscious remedy must be closely tied to the relevant labor market. See Paradise, 480 U.S. at 179. To weigh this factor, common sense demands we first know the remedy’s numerical goal and the relevant labor market. As discussed above, we know neither.

Croson mandates the number of minorities in the relevant labor market be determined by the “number of minorities qualified to undertake the particular task.” Croson, 488 U.S. at 502. As we pointed out first in our compelling interest analysis, and again in the necessity and efficacy portion of our narrow tailoring analysis, the City has not provided us with the demographics of its *25 qualified applicant pool. Therefore, we cannot properly define the “relevant labor market.” This alone prevents us from reviewing and weighing this factor. Further, as discussed in our duration analysis, the long-term goal of the decree is unclear. Until we know the numerical goal of the decree and the relevant labor market, the relationship between the numerical goal and the relevant labor market remains a genuine issue of material fact to be resolved on remand.

d. Impact of the Relief on Third Parties

The final Paradise factor ensures a remedy does “not impose an unacceptable burden on innocent third parties.” Paradise, 480 U.S. at 182. The district court found the impact on third parties “not overly significant” in this case. See Dean, slip op. at 18. The Supreme Court has given little guidance on this factor, but has made a few things clear. First, remedies requiring nonminorities to be fired impose a severe, and possibly unacceptable, burden on third parties. Paradise, 480 U.S. at 182. Hiring preferences are less burdensome. See id. Second, remedies allowing unqualified minorities to be hired are likely not narrowly tailored. See id. at 183. Third, remedies merely postponing a benefit to third parties are less burdensome than ones permanently denying a benefit. See id.

We agree with the district court that the impact on nonminorities is not significant enough to make the decree and *26 hiring policy unconstitutional per se. We do so because the remedies in this case have all three characteristics the Supreme Court has previously said favor a finding of narrow tailoring within the context of this factor: (1) they do not require nonminorities to be fired; (2) they do not require unqualified minorities to be hired; and (3) they do not pose an absolute bar to nonminority employment. Paradise, 480 U.S. at 182-83. This factor alone does not prevent a finding of narrow tailoring.

In sum, the district court on remand must develop the record further and re-evaluate both whether the decree and the hiring process were still justified at the time of suit by a compelling government interest and whether they were narrowly tailored to further that interest. III. Appellants’ Title VII Claim

Appellants also argue that the City’s hiring process violates Title VII. Specifically, Appellants refer us to 42 U.S.C. § 2000e-2(l), which was added to Title VII in 1991 and provides that “[i]t shall be an unlawful employment practice . . . in connection with the selection or referral of applicants or candidates for employment . . . to adjust the scores of, use different cutoff scores for, or otherwise alter the results of, employment related tests on the basis of race, color, religion, sex, or national origin.” § 2000e-2(l) (emphasis added) Appellants claim that by separating applicants’ Civil Service Exam scores by race, the City *27 in effect uses different cutoff scores on the basis of race. We agree that the City’s hiring process violates the plain language of section 2000e-2(l).

First, we must point out that our decision that the City’s hiring process violates section 2000e-2(l) will stand even if on remand the district court finds that the hiring process passes scrutiny under the Equal Protection Clause. That is, even if the process is “upheld as a valid method of affirmative action,” Chicago Firefighters Local 2 v. City of Chicago, 249 F.3d 649, 656 (7th Cir. 2001), it nevertheless violates the specific prohibitions of section 2000e-2(l). See id. (recognizing that the practice of “banding” test scores is an acceptable form of affirmative action under other Title VII provisions and the Equal Protection Clause, but continuing on to determine whether that practice violated section 2000e-2(l)).

Appellants insist that the City violates section 2000e-2(l) when, pursuant to its hiring policy, it ranks test takers from highest to lowest exam score, but then separates those scores according to race and sex. The City responds, and the district court held, that the hiring process does not violate the statute because “[e]very applicant, black or white, ha[s] to score at least a seventy-five to pass the Civil Service Exam.” Dean, slip op. at 20. This conclusion is incomplete, however, because it focuses only on the initial portion of phase one of the City’s hiring *28 process instead of on the entire process. The City is correct that all applicants are subject to the same initial requirement at phase one -- a score of at least seventy-five on the Civil Service Exam. Our analysis, however, cannot stop there because the City uses the test scores again at another important step in the hiring process. A passing score of seventy-five only makes applicants eligible for further consideration. Later in the process the City again uses the exam scores to choose which applicants will proceed to phase two of the hiring process. The City separates applicants’ exam scores by race and sex and selects the same number of blacks and whites to proceed, starting with the highest exam score on each segregated list.

This method of selection between phase one and phase two violates the plain language of section 2000e-2(l) because it has the practical effect of requiring different cutoff scores, based solely on race and sex, for continuing further in the hiring process. We see no reason to ignore a clear violation of section 2000e-2(l) between phase one and phase two simply because at the outset of phase one the same cutoff score is required of all applicants. Compliance with section 2000e-2(l) is required at all times during employment activities carried out “in connection with the selection or referral of applicants or candidates for employment.” See 42 U.S.C. § 2000e-2(l).

We find that the City’s hiring process violates Title VII. *29 Therefore, we reverse the district court’s dismissal of Appellants’ Title VII claim. IV. Appellants’ Louisiana Constitutional Claim

Appellants’ next contention is that the City’s hiring process violates the Louisiana Constitution. Appellants point to Article I, Section 3, which states not only that “[n]o person shall be denied the equal protection of the Laws,” but also that “[n]o law shall discriminate against a person because of race.” The district court, after determining that the City’s hiring process survived scrutiny under the United States Constitution, held that regardless of whether the hiring process violates the Louisiana Constitution, “Louisiana law must bow to federal law, which is the supreme law of the land.” Dean, slip op. at 21. On appeal, the City simply states that even if its hiring process violates the Louisiana Constitution, the consent decree shields the City from liability because it “preempts state law.” We disagree and find that the district court erred in determining that preemption affords the City any form of protection from possible violations of state law.

Clearly, had the City been required by federal law to implement its hiring process to redress a past federal constitutional violation, federal law would preempt the Louisiana Constitution. See In re Birmingham Reverse Discrimination Employment Litig., 833 F.2d 1492, 1501 (11th Cir. 1987). But that is not the situation we face here. The City originally denied that *30 it had engaged in intentional discrimination and was not found to be in violation of the United States Constitution. More importantly, the City was never ordered by a federal court, nor mandated by any other federal law, to enter into the consent decree. Rather, the City did so voluntarily. Thus, in the eyes of the law the City’s hiring process is nothing more than a voluntary affirmative action program.

As the Eleventh Circuit has stated:

We perceive no reason for treating a consent decree entered pursuant to a voluntary settlement differently from a voluntary affirmative action plan. In both instances, the employer has embarked on a voluntary undertaking; we reject any notion that the memorialization of that voluntary undertaking in the form of a consent decree somehow provides the employer with extra protection against charges of illegal discrimination . . . . It should be emphasized that there has been no judicial determination that the City is liable for past discrimination . . . . Thus . . . we are not presented with a case in which the defendant was required by law to implement an affirmative action program designed to remedy the effects of past discrimination.

In re Birmingham Reverse Discrimination Employment Litig., 833 F.2d at 1501 (11th Cir. 1987). We think the Eleventh Circuit’s reasoning is sound. As far as preemption is concerned, a voluntary consent decree has the same effect on state law as does a voluntary affirmative action program--none.

Therefore, we need only decide whether the City’s practice of separating applicants by race during its hiring process violates the Louisiana Constitution. We think it is beyond dispute that it *31 does. Decisions of the Supreme Court of Louisiana bind us with regard to the meaning of Louisiana constitutional provisions. Those decisions make abundantly clear that Article I, Section 3 of the Louisiana Constitution provides far greater protection against racial discrimination than does its federal counterpart. See, e.g., La. Associated Gen. Contractors, Inc. v. State, 669 So. 2d 1185, 1196 (La. 1996). Under the United States Constitution, classifications based on race are permissible if they are narrowly tailored to serve a compelling government interest. See Paradise, 480 U.S. at 166-67. However, under the Louisiana Constitution, classifications based on race “shall be repudiated completely, regardless of the justification.” La. Associated Gen. Contractors, Inc., 669 So. 2d at 1198. Under Louisiana law, once it is determined that a classification based on race has been drawn, the inquiry is over--Article I, Section 3 of the Louisiana Constitution has been violated. See id. Here, the City’s hiring process unquestionably classifies according to race. The City separates white and black firefighter applicants when deciding which applicants will proceed to phase two of the hiring process. The City’s actions violate Article I, Section 3 of the Louisiana Constitution. Also, even if on remand the district court determines that the consent decree survives strict scrutiny under the United States Constitution, an outcome which is far from certain, the Louisiana Constitution is not preempted because the *32 decree was entered into voluntarily. Therefore, we reverse the district court’s dismissal of Appellants’ Louisiana constitutional claim. V. Appellants’ Louisiana Statutory Claim

Appellants’ final argument is that the City’s hiring process violates a Louisiana anti-discrimination employment statute. Appellants note that LA. REV. STAT. ANN. § 23:332(A)(2) (1998) makes it unlawful for employers in Louisiana to “[i]ntentionally limit, segregate, or classify . . . employees or applicants . . . in any way which would deprive or tend to deprive any individual of employment opportunities, or otherwise adversely affect his status as an employee, because of the individual’s race, color, religion, sex, or national origin.”

The district court granted summary judgment in favor of the City dismissing this claim. We agree with the court’s ruling on this claim because Appellants’ argument is foreclosed by the plain language of another subsection of the statute they rely upon, which provides that “[n]othing contained in [section 23:332] shall be construed so as to create a cause of action against an employer . . . for employment practices pursuant to any affirmative action plan.” LA. REV. STAT. ANN. § 23:332(G) (emphasis added). The statute make no distinction between valid and invalid affirmative action plans. See id. Therefore, regardless of the outcome on remand of Appellants’ remaining claim, the City’s practice of *33 classifying applicants according to race and sex does not violate this statute. We affirm the district court’s dismissal of Appellants’ Louisiana statutory claim.

CONCLUSION

Based upon the foregoing, we AFFIRM the district court’s dismissal of Appellants’ Louisiana statutory claim; we REVERSE its dismissal of Appellants’ equal protection claim and REMAND the case for further factual development of the equal protection claim; and we REVERSE its dismissal of Appellants’ Title VII and Louisiana constitutional claims, and with respect to the Title VII and Louisiana constitutional claims, we REMAND the case for further proceedings concerning the entitlement of each individual Appellant to relief. AFFIRMED in part; REVERSED in part; and REMANDED for further proceedings consistent with this opinion.