The opinion of the court was delivered by: Nicholas G. Garaufis, United States District Judge.
In this Opinion, the court considers Plaintiffs' Motions for Summary Judgment for backpay damages and several other issues relating to the individual claims process. Upon due consideration, the court denies Plaintiffs' Motions but does find there is no genuine issue as to the amount of aggregate pre-mitigation wage backpay owed through 2010. The court grants the City leave to amend its Answer to raise the defense of failure to mitigate. The court grants the parties' joint request for a definition of Non-Hire Claimant and Delayed-Hire Claimant, and the use of six additional criteria agreed upon by the parties. The court denies the City's request for the use of criteria beyond those agreed to by all parties. The court selects four special masters and directs them to file affidavits pursuant to Federal Rule of Civil Procedure 53(3)(A-B). The court denies the City's motion to delay the individual determinations of monetary relief during the pendency of the City's appeal of the court's Remedial Order. Finally, the court orders the parties to appear for a litigation management conference on March 22, 2012, at 1:00 p.m.
The factual and procedural background of this case is complex, and will not be recounted here in full. Only the events that are relevant to the issues before the court today will be summarized.*fn1
In 2007, the United States brought suit against the City of New York (the "City"), alleging that certain aspects of the City's policies for selecting entry-level firefighters for the New York City Fire Department ("FDNY") violated Title VII of the 1964 Civil Rights Act, as amended, 42 U.S.C. § 2000e et seq. ("Title VII"). (July 22, 2009 Mem. & Order (Docket Entry # 294) (the "D.I. Op.") at 2.) The United States alleged that the City's use of two exams- Written Exams 7029 and 2043- as pass-fail screening and rank-ordering devices had a disparate impact on black and Hispanic candidates for entry-level firefighter. (Id. at 3.) The Vulcan Society and several individuals intervened in the lawsuit as plaintiffs, alleging similar claims of disparate impact and also alleging disparate treatment (raising both theories of liability under federal, state, and local law) on behalf of a class of black entry-level firefighter candidates. (See Sept. 5, 2007 Mem. & Order (Docket Entry # 47) at 20 (granting motion to intervene); May 11, 2009 Mem. & Order (Docket Entry # 281) at 33-34 (certifying a class for the liability stage of the case).)
In July 2009, the court considered the United States' and Plaintiff-Intervenors' Motions for Summary Judgment on disparate impact liability. (D.I. Op. at 93.) In its opinion, the court recounted the City's basic process for selecting entry-level firefighters. Applicants applied to sit for a written, multiple choice exam; the City administered Exam 7029 from 1999 until 2002 and Exam 2043 from 2002 until 2007. (Id. at 10.) Each exam had a pass-fail cutoff; applicants who scored at or above that cutoff were allowed to take a physical performance test (the "PPT"), while those who failed were not allowed to take the PPT or progress any further in the hiring process. (Id.) Candidates who also passed the PPT were placed on a rank-order hiring eligibility list. (Id. at 11.) Applicants' list rank was determined by a final score that was produced through a process of combining applicants' written exam and PPT scores (along with additional points for New York City Residents, legacy candidates, and veterans). (Id.) When the FDNY intended to hire entry-level firefighters for a new class, the City would certify a portion of the eligibility list (by rank order) for further processing such as background checks and medical and psychological examinations; a person who passed through all of the further processing was appointed. (Id. at 11-12.) The City used the eligibility list from Exam 7029 to hire firefighter classes from 2001 through 2004, and the list from Exam 2043 from 2004 through 2008. (Id. at 12.)
After recounting the basic FDNY firefighter hiring process, the court considered the evidence of a prima facie case of disparate impact. Plaintiffs challenged two practices for each of the two exams: the pass-fail cutoff score and the rank-ordering of candidates based on their scores. (Id. at 24.) The statistical evidence Plaintiffs offered (largely though not exclusively through the work of the United States' expert, Bernard R. Siskin, Ph.D.) was telling.
As to the pass-fail usage of the exams, blacks passed Exam 7029 at 67% of the pass rate of whites; the disparity was equivalent to 33.9 units of standard deviation, meaning the likelihood of the disparity occurring by chance was "infinitesimally small."*fn2 (Siskin Liability Report (Docket Entry # 254 Ex. A) at 3, 21, & Table 1.) Hispanics passed Exam 7029 at 85.3% of the pass rate of whites; the disparity was equivalent to 17.4 units of standard deviation, meaning the likelihood of the disparity occurring by chance was infinitesimally small. (Id. at 3, 23, & Table 2.) Blacks passed Exam 2043 at 87.7% of the pass rate of whites; the disparity was equivalent to 21.8 units of standard deviation, meaning the likelihood of the disparity occurring by chance was infinitesimally small. (Id. at 5, 26, & Table 5.) Hispanics passed Exam 2043 at 95.5% of the pass rate of whites; the disparity was equivalent to 10.5 units of standard deviation, meaning the likelihood of the disparity occurring by chance was infinitesimally small. (Id. at 5-6, 27-28, & Table 6.)
As to the use of the exams as rank-ordering devices, blacks were disproportionately ranked disadvantageously on the Exam 7029 list; for example, while 33% of white candidates had eligibility list numbers at or above 2000, only 21% of black candidates did, and while 20% of white candidates had list numbers at or below 5001, 30% of black candidates did. (United States Rule Disparate Impact Motion 56.1 Statement (Docket Entry # 252) ¶¶ 118-19.) The disparity in ranking is equivalent to 6.48 units of standard deviation, meaning the likelihood of the disparity occurring by chance was less than 1 in 11 billion. (Siskin Liability Report at 4 & 24.) Similarly, the eligibility list created from the Exam 7029 results disproportionately ranked Hispanic candidates disadvantageously. While 33% of white candidates had eligibility list numbers at or above 2000, only 28% of Hispanic candidates did, and while 20% of white candidates had list numbers at or below 5001, 29% of Hispanic candidates did. (United States Rule 56.1 Statement ¶¶ 128-29.) The disparity in ranking was equivalent to 4.57 units of standard deviation, meaning the likelihood of the disparity occurring by chance was less than 1 in 204,000. (Siskin Liability Report at 4 & 25.) Blacks were disproportionately ranked disadvantageously on the Exam 2043 eligibility list. While 28% of white candidates had eligibility list numbers at or above 2000, only 18% of black candidates did, and while 30% of white candidates had list numbers at or below 5001, 50% of black candidates did. (Id. ¶¶ 137-39.) The disparity in ranking was equivalent to 9.5 units of standard deviation, meaning the likelihood of the disparity occurring by chance was calculated to be less than 1 in 4.5 million-billion. (Siskin Liability Report at 7 & 32.) Hispanics were disproportionately ranked disadvantageously on the Exam 2043 eligibility list. While 28% of white candidates had eligibility list numbers at or above 2000, only 25% of Hispanic candidates did, and while 30% of white candidates had list numbers at or below 5001, 39% of Hispanic candidates did. (United States Rule 56.1 Statement ¶¶ 150-51.) The disparity in ranking was equivalent to 4.6 units of standard deviation, meaning the likelihood of the disparity occurring by chance was less than 1 in 186,225. (Siskin Liability Report at 8 & 34.)
Plaintiffs put forward evidence about the practical significance of the statistical disparities documented above. As to the pass-fail use of the exams, the disparity in the rate at which black candidates passed Exam 7029 accounted for 74.7% of failing black candidates- eliminating 519 candidates from contention (these candidates are called "shortfall candidates").
From those 519, approximately 114 additional black entry-level firefighters would have been selected ("shortfall hires"). (Id. at 3, 22, & Table 1.) Dr. Siskin calculated the approximate number of shortfall hires by assuming that the shortfall candidates would have passed the PPT and completed the other hiring requirements at the same rate as passing candidates did. (Id. at 16-17.) The disparity in the rate at which Hispanic candidates passed Exam 7029 accounted for 56.9% of failing Hispanic candidates-eliminating 282 candidates from consideration. From those 282, 62 additional Hispanic entry-level firefighters would have been appointed. (Id. at 3, 23-24, & Table 2.) As for Exam 2043, the disparity in the rate at which black candidates failed the Exam accounted for 81.3% of the black candidates who failed-eliminating 165 black candidates from consideration. From those 165 candidates, 30 additional black firefighters would have been hired. (Id. at 5-6, 27, & Table 5.) The disparity in the rate at which Hispanic candidates failed Exam 2043 accounted for 61.8% of the Hispanic candidates who failed- eliminating 94 Hispanic candidates from consideration. From those 94 candidates, an estimated 17 would have been hired. (Id. at 6, 28, & Table 6.)
Plaintiffs put forward evidence of the practical significance of the rank-order usage of the Exams as well. For example, Plaintiffs offered statistical evidence that the disproportionately low ranking of black candidates on the 7029 list meant that out of 104 black candidates who were appointed to be entry-level firefighters, 68 were appointed disproportionately late, creating an aggregate total of 20 years of lost wages and seniority. (Id. at 4, 25, & Table 3 Part B.) Similarly, the disproportionately low average location of Hispanic candidates on the 7029 list meant that out of 274 Hispanic candidates who were hired, 86 were delayed in their appointments, creating an aggregate total of approximately 23 years of lost wages and seniority. (Id. at 4-5, 26, & Table 4 Part B.)
Plaintiffs offered evidence of the practical significance of the rank-order usage of Exam 2043 as well. For example, the disproportionately low rank of the 80 black candidates on the 2043 list who were hired meant that 44 were delayed in their appointments, for an aggregate estimated total of 14 years of delayed wages and seniority. (Id. at 9, 33-34, & Table 12 Part B.) Similarly, the disproportionately low rank of the Hispanic candidates on the 2043 list meant that 51 out of the 187 Hispanic hires off the list were delayed in appointment, resulting in an aggregate total of 12 years worth of delayed wages and seniority. (Id. at 9, 35, & Table 14 Part B.)
However, the eligibility list produced from Exam 2043 differed from that produced from Exam 7029 in at least one important way: the 2043 list was not exhausted, i.e., the City did not certify every candidate who had passed Exam 2043 and the PPT; the 7029 list, on the other hand, was exhausted eventually. (See id. at 18 n.9.) Consequently, the practical significance of the rank-order usage of Exam 2043 took an additional form. As described by Dr. Siskin, candidates who ranked too poorly on the list (i.e., had combined scores that were relatively low) had "failed" the "rankings test" of the Exam. (Id. at 32-35.) Dr. Siskin generated a statistical analysis of the disparity between the number of white candidates who had passed the rankings test by achieving a rank high enough to be considered for hiring and the black and Hispanic candidates who had been similarly ranked. Black candidates were disproportionately likely to have been ranked too poorly to have been considered for hiring. This disparity produced a shortfall of 95 candidates, and out of those 95 candidates, 42 would have been hired as entry-level firefighters. (Id. at 19, 32-33, & Table 11.) Similarly, Hispanic candidates were disproportionately likely to have been ranked too poorly to have ever been considered for hiring.
This disparity produced a shortfall of 63 candidates, and out of those 63 candidates, an estimated 28 additional Hispanic firefighters would have been hired. (Id. at 19, 34-35, & Table 13.)
The court concluded that the undisputed statistical evidence Plaintiffs put forward showed that black and Hispanic candidates disproportionately failed Written Exams 7029 and 2043, were ranked disproportionately lower on the eligibility lists the exams created, and that the standard deviation analysis offered showed a very small likelihood that any of the disparities occurred by chance. (D.I. Op. at 26.) The standard deviations calculated "strongly support[ed] a conclusion of a causal relationship between the observed disparities and the employment practices at issue." (Id.) The court concluded that the significance of the statistical evidence of disparity was bolstered by the evidence of practical significance. The court was persuaded by the United States' expert's analysis of the practical significance of each of the disparities:
[A]pproximately one thousand additional black and Hispanic candidates would have been considered for appointment as FDNY firefighters had it not been for the disparities resulting from the examinations. Further, absent these disparities, approximately 293 additional black and Hispanic candidates would have been appointed from the eligibility lists used from 2001 through 2008, and approximately 249 black and Hispanic applicants who were actually appointed would have been appointed sooner. . . . [I]t is clear that these disparities have had a practical significance. (Id. at 27.) The court concluded that the evidence for a prima facie case of disparate impact was "overwhelming." (Id.)
The court then considered the City's arguments in opposition to Plaintiffs' prima facie case of disparate impact. The City had put forward three arguments: that the "80% rule" should be used instead of statistical significance analysis (id. at 33-34); that the size of the populations tested in the case was too large to produce statistical significance tests (id. at 28-31); and that Dr. Siskin's "shortfall" analysis improperly assumed that all races and ethnic groups could perform equally well on the exams and other components of the hiring process (id. at 32-33). The court rejected each these arguments as nothing more than metaphysical doubts about the nature of statistical evidence that were unsupported in law. (Id. at 35.) The court granted summary judgment to Plaintiffs on the prima facie case of disparate impact: "[t]here is no dispute that Plaintiffs have satisfied a statistical standard for a prima facie case of disparate impact . . . , nor is there any dispute that they have shown that the disparity has had a substantial, practical significance for the composition of the eligibility lists and hiring of entry-level firefighters." (Id.) The court then considered the City's business necessity defense, and after an exhaustive review of whether the process that the City had used to create the challenged exams was sufficient to satisfy Second Circuit precedent on the issue of the business necessity of employment examinations, concluded that the City had offered no genuine issue of fact as to the business necessity defense. (Id. at 50-91.)
The court later considered Plaintiff-Intervenors' Motion for Summary Judgment on Disparate Treatment Liability. As relevant to the issues before the court in the instant opinion, the court cited much of the statistical evidence discussed above as part of the prima facie case of disparate treatment discrimination by the City against black firefighters. (See Jan. 13, 2010 Mem. & Order (Docket Entry # 385) ("D.T. Op.") at 29-31.) The court granted summary judgment to Plaintiff-Intervenors as to the City's liability on Plaintiff-Intervenors' claims of disparate treatment, violation of the Equal Protection clause, and state and local law disparate impact and disparate treatment claims. (See id. at 70.)
After the court issued both of its liability opinions, it proceeded to the remedial phase of the case. The court issued an Initial Remedial Order which provided a preliminary outline for the remedial phase. As is relevant for the purposes of today's opinion, the court made clear that it would order compensatory, "make whole" relief for the individual victims of the disparate impact of the entry-level firefighter hiring process. (Jan. 21, 2010 Mem. & Order (Docket Entry # 390) (the "I.R. Order") at 10.) Following the Supreme Court's seminal opinion in Franks v. Bowman Transportation, 424 U.S. 747 (1976), the court ruled that compensatory relief would be available to actual test-takers-any black or Hispanic person who sat for Exams 7029 or 2043- unless the City could prove that particular individual applicants would not have been hired for non-discriminatory, lawful reasons. (I.R. Order at 13-15.) A claimant's initial burden would be to show that he or she took an exam and was not hired. (Id.) The court noted that the initial remedial plan the United States had proposed gave the United States the initial authority to determine a claimant's eligibility. The court noted that a special master might do the same job. (Id. at 16-17 & n.11.) The court encouraged the parties to develop an agreed-upon list of criteria that would disqualify claimants. (Id. at 17-18.) The court next reviewed the types of relief potentially available for claimants-monetary and hiring-and the primary difference in eligibility for the two types: a claimant would be eligible for monetary relief if he or she was eligible to be hired at the time the exams were administered, but only eligible for hiring relief if currently eligible to be hired according to the City's non-discriminatory hiring criteria. (Id. at 19-20.) The court left open what exact criteria a claimant might have to demonstrate to be eligible for priority hiring relief, and how the parties would resolve who would receive hiring relief if more than 293 individuals were eligible and interested. (Id. at 20-22 n.18.) The court's Initial Remedial Order operated on the underlying principle that the City should hire 293 candidates from the victims as part of the make-whole relief required by the City's prior conduct. (Id.) The court further considered the parties' suggestions on whether to award retroactive seniority to the 293 priority hires; the City objected to awarding retroactive seniority, but was silent on the number of priority hires. (Id. at 23-24.) The court overruled the City's objection and reiterated that priority hiring would be limited to 293 positions because that was the shortfall number determined in the disparate impact liability opinion. (Id. at 25-27.) The court noted that monetary relief would be based on the same shortfall number but distributed pro rata to the larger pool of eligible claimants. (Id. 25 n.22.) The court then expressed intent to order retroactive seniority for claimants who were hired by the FDNY but whose hiring was delayed. The court relied on the statistical analysis that it had approved in the Disparate Impact Liability Opinion for the number of delayed hires for each ethnic group and each exam, and for the number of years of wages and seniority lost due to the delays. (Id. at 29.) The court ordered the parties to consider various issues related to creating a notice-and-claims process for compensatory relief. (Id. at 56.)
Most recently, the court considered Plaintiff-Intervenors' motion for class certification in the remedial phase. (June 6, 2011 Mem. & Order (Docket Entry # 640) at 9-18.) While the court conditionally certified subclasses (the Delayed-Hire Victim Subclass and the Non-Hire Victim Subclass) to allow for class-wide resolution of several issues of make-whole relief (backpay, benefits, and retroactive seniority and priority hiring (see id. at 14)), the court refused to certify any subclass with regard to the issue of mitigation of damages (id. at 24-25). The court reasoned that resolution of mitigation of damages will turn on individual issues-the earnings and actions of each class member-and therefore is not an issue common to the class. (Id. at 21-22.) The court further observed that class-wide resolution of mitigation is unnecessary because the parties had already planned for a notice-and-claims process for individual claimants. The court proposed modifying that process to include a process for considering the extent to which each claimant mitigated his or her damages. (Id. at 24.) The court observed that this same logic, which compelled the court to deny class certification as to mitigation, would also compel it to deny summary judgment as to a class-wide mitigation calculation. (Id. at 25 n.12.)
The court also concluded that exceptional conditions justified the appointment of a special master to conduct an individual claims process under Federal Rule of Civil Procedure 53(a)(1)(B). (Id. at 40.) The court proposed that special masters would be used to make findings of fact and conclusions of law as to individual claimants on certain issues: (1) each claimant's eligibility to receive individual relief; (2) the share of the aggregate amount of backpay to which each eligible claimant is entitled; (3) the extent of a claimant's mitigation and, if the City meets its burden of production on this question, the reasonableness of a claimant's mitigation efforts and the availability of substitute employment; (4) for black claimants, the amount of the claimant's compensatory damages for non-economic losses; and (5) the claimant's eligibility for priority-hiring relief and/or the amount of retroactive seniority, if applicable. (Id.) The court also projected a need for the special masters to oversee settlement negotiations and a discovery process to collect information from individual claimants. (Id.) The court suggested that the role of special master might be filled by a United States Magistrate Judge appointed pursuant to Rule 53(h). (Id. at 41.) However, the court did not make a final determination on that point, nor did it appoint a special master.
II.PLAINTIFFS' MOTION FOR SUMMARY JUDGMENT FOR BACKPAY
1. Summary Judgment Standard
A motion for summary judgment must be granted if "the movant shows that there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law." Fed. R. Civ. P. 56(a). In determining whether a genuine issue of material fact exists, "the court must draw all reasonable inferences in favor of the nonmoving party." Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 149 (2000). Further, the burden of showing the absence of any genuine dispute as to a material fact rests on the movant. See Adickes v. S.H. Kress & Co., 398 U.S. 144, 157 (1970).A fact is material if its existence or non-existence "might affect the outcome of the suit under the governing law," and an issue of fact is genuine if "the evidence is such that a reasonable jury could return a verdict for the nonmoving party." Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986). Rule 56 "mandates the entry of summary judgment . . . against a party who fails to make a showing sufficient to establish the existence of an element essential to that party's case, and on which that party will bear the burden of proof at trial." Celotex Corp. v. Catrett, 477 U.S. 317, 322 (1986). In such a situation, "there can be 'no genuine issue as to any material fact,' since a complete failure of proof concerning an essential element of the nonmoving party's case necessarily renders all other facts immaterial." Id. at 323. A grant of summary judgment is proper "[w]hen no rational jury could find in favor of the nonmoving party because the evidence to support its case is so slight." Gallo v. Prudential Res. Servs., L.P., 22 F.3d 1219, 1224 (2d Cir. 1994).
The party opposing summary judgment is not entitled to rely on unsworn allegations in the pleadings, but must instead "show that there is admissible evidence sufficient to support a finding in her favor on the issue that is the basis for the motion." Fitzgerald v. Henderson, 251 F.3d 345, 360-61 (2d Cir. 2001). A "genuine issue [is not] created merely by the presentation of assertions that are conclusory." Patterson v. County of Onieda, N.Y., 375 F.3d 206, 219 (2d Cir. 2004). Likewise "conjecture or speculation by the party resisting the motion will not defeat summary judgment." Kulak v. City of New York, 88 F.3d 63, 71 (2d Cir. 1996).
2. Title VII Backpay Standards
The legal principles relevant to Plaintiffs' Summary Judgment Motion are for the most part pellucid. Municipal employers are subject to Title VII's prohibition on employment discrimination. See Annis v. County of Westchester, 36 F.3d 251, 254 (2d Cir. 1994). Disparate impact has been a recognized theory of liability under Title VII since Griggs v. Duke Power Co., 401 U.S. 424, 431 (1971). Where a court has found a violation of Title VII, it may consider a number of equitable remedies: "the court may . . . order such [relief] as may be appropriate, which may include, but is not limited to, reinstatement or hiring of employees, with or without back pay, . . . or any other equitable relief as the court deems appropriate." 42 U.S.C. § 2000e-5(g)(1). The purpose of providing the court with equitable authority in Title VII cases is to "allow the most complete achievement of the objectives of Title VII that is attainable under the facts and circumstances of the case." Franks v. Bowman Transportation Co., 424 U.S. 747, 770-71 (1976). One objective of Title VII is to "make persons whole" if they have suffered employment discrimination. Int'l Bhd. of Teamsters v. United States, 431 U.S. 324, 364 (1977).
The Supreme Court has made clear that the use of the words "equitable" and "may" in 42 U.S.C. § 2000e-5(g)(1) do not imply that district courts have unfettered discretion when considering whether to grant an award of backpay. See Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975). Considering the purposes Congress contemplated when it drafted Title VII, the Court held that "given a finding of unlawful discrimination, backpay should be denied only for reasons which, if applied generally, would not frustrate the central statutory purpose of eradicating discrimination throughout the economy and making persons whole for injuries suffered through past discrimination." Id. at 421. Backpay is one of the basic components of "make whole" relief. Wrenn v. Dep't of Veterans Affairs, 918 F.2d 1073, 1076 (2d Cir. 1990). Backpay consists of "lost salary, . . . anticipated raises, and fringe benefits." Saulpaugh v. Monroe Community Hosp., 4 F.3d 134, 145 (2d Cir. 1993). In addition to those principal components of backpay, "it is ordinarily an abuse of discretion not to include pre-judgment interest in a back-pay award." Id. (quoting Clarke v. Frank, 960 F.2d 1146, 1153-54 (2d Cir. 1992)) (emphasis in original). In fashioning a backpay award, the court "must, as nearly as possible, recreate the conditions and relationships that would have been had there been no unlawful discrimination." Ingram v. Madison Square Garden, Inc., 709 F.2d 807, 811 (1983) (internal quotation marks removed). Backpay can be awarded in class actions; however, where the facts are not so clear as to allow a determination as to which class members should be awarded backpay, a court may equitably "compute a gross award for all the injured class members and divide it among them on a pro rata basis." Id. at 812. For example, "where the number of qualified class members exceeds the number of openings lost to the class through discrimination and identification of individuals entitled to relief would drag the court into a quagmire of hypothetical judgments and result in mere guesswork," Catlett v. Mo. Highway & Transp. Comm'n, 828 F.2d 1260, 1267 (8th Cir. 1987), a case "may require class-wide, rather than individualized, assessments of monetary relief," Robinson v. Metro-North Commuter R.R., 267 F.3d 147, 161 n.6 (2d Cir. 2001), overruled in part, Wal-Mart v. Dukes, No. 10-277, 2011 WL 2437013, at *12-15 (U.S. June 20, 2011).
Employers' liability for backpay extends to two years before the filing of the relevant EEOC charge. 42 U.S.C. § 2000e-5(g)(1). However, in a class action, unnamed class members do not have to file a charge with the EEOC to be eligible for backpay. Albemarle Paper Co., 422 U.S. at 414 n.8.
While it is clear that backpay includes the value of fringe benefits, see Saulpaugh, 4 F.3d at 145, the law is less clear with regard to how to value some of those benefits, such as employer-provided health insurance. Some courts have held that an employer is liable for the amount it would have paid in premiums for an employee's health insurance. See, e.g., EEOC v. Dial Corp., 469 F.3d 735, 744 (8th Cir. 2006); Mister v. Illinois Cent. Gulf R.R., 790 F. Supp. 1411, 1418-19 (S.D. Ill. 1992) (applying this method of calculating value in a class action context). However, the weight of authority appears to be in favor of a contrary rule-that an employer is liable for an employee's out-of-pocket expenses that would have been covered under the employer's health plan. See, e.g., Galindo v. Stoody Co., 793 F.2d 1502, 1517 (9th Cir. 1986) (interpreting backpay provisions of the National Labor Relations Act and holding that "plaintiff should be compensated for the loss of [health] benefits if the plaintiff has purchased substitute insurance coverage or has incurred, uninsured, out-of-pocket medical expenses for which he or she would have been reimbursed"); Mugavero v. Arms Acres, Inc., 689 F. Supp. 2d 544, 582-83 (S.D.N.Y. 2010) (in Title VII case, requiring proof of what benefits plaintiff received at previous employer and thus would be entitled to reimbursement for out-of-pocket expenses that would have been covered); Evans v. State of Conn., 967 F. Supp. 673, 683 (D. Conn. 1997) (in Title VII case, "plaintiff can only be compensated for his actual out of pocket cost of medical insurance"). As the court discusses in greater detail below, the court concludes that the latter approach is the appropriate one.
The United States bases its Motion for Summary Judgment largely on the statistical analyses of Dr. Siskin. Dr. Siskin's work involved six steps: first, estimating the hiring shortfall and the academy class into which each shortfall candidate should have been placed; second, estimating the wages that the shortfall hires would have earned had they been hired between their date of hire and the point at which losses stop accruing; third, determining the likely rate of attrition for each shortfall hire; next, estimating the earnings that the shortfall hires would have been reasonably likely to have made in a job other than that of firefighter, and reducing the total estimate of backpay accordingly; fifth, accounting for the value of fringe benefits shortfall hires would have earned as firefighters; and last, applying a rate of interest to the estimates. (Siskin Relief Phase Report (Docket Entry # 538-2) at 4.) The City objects to the methodology and results for some of the steps. (City Mem. in Opp. to Backpay Mot. Summ. J. (Docket Entry # 543) at 2.) The court will review each of Dr. Siskin's steps along with the City's objections, if any, sequentially. The court first discusses the methodology and process for determining the loss of non-hire victims and then the same inquiry for the loss of delayed-hire victims.
Dr. Siskin's first step was determining the hiring shortfall and assigning shortfall hires to academy classes. (Siskin Relief Phase Report at 4.) Dr. Siskin's process for determining the shortfall hiring number was simple: he used the shortfall numbers that the court used in its Disparate Impact Liability Opinion, which in turn the court had drawn from Siskin's previous report. (Id. at 4-5.) The hiring shortfalls that the court cited in that opinion were: a shortfall of 114 black firefighters from Exam 7029; a shortfall of 62 Hispanic firefighters from Exam 7029; a shortfall of 72 black firefighters from Exam 2043 (30 due to the pass-fail use of the Exam and 42 due to the rank-order use of the Exam); and a shortfall 45 Hispanic firefighters from Exam 2043 (17 from the pass-fail use of the Exam and 28 due to the rank-order use of the Exam). See Section I, supra.
Dr. Siskin next estimated how the shortfall hires would have been distributed to different academy classes in the absence of discrimination. (Siskin Relief Phase Report at 5.) Dr. Siskin allocated the shortfall hires to academy classes based on the percentage of the total list that was inducted into each class; e.g., if the first academy class hired off of the 7029 list took 1% of the total hires from that list, then Dr. Siskin allocated 1% of the black and Hispanic shortfall candidates from the 7029 Exam list to that class. (Id.) Dr. Siskin's data for the total number and distribution across academy classes of actual hires came from information produced by the City in discovery. (Id.)
A tabular representation of Dr. Siskin's distribution of the shortfall from the 7029 Exam follows: 7029 List Academy Classes and Shortfall Hires
Academy Class Actual Number of Actual Hire Black shortfall hires Hispanic shortfall
Start Date Firefighters Hired Percentage assigned hires assigned
9.12.2004 174 5.4% 6 3 Total 3,202 100.0% 114 62
(See Siskin Relief Phase Report Table A-1.)
A tabular representation of Dr. Siskin's distribution of the shortfall from the 2043 Exam follows: 2043 List Academy Classes and Shortfall Hires
Academy Class Actual Number of Actual Hire Black shortfall hires Hispanic shortfall Start Date Firefighters Hired Percentage assigned hires assigned
8.5.2007 289 13.5% 10 6 Total 2148 100.0% 72 45 (See id.)
The City objects to Dr. Siskin's estimates on several grounds. The
City's relief phase expert, Dr. Christopher Erath, echoes some of
these objections. The first objection is that Dr. Siskin's shortfall
calculation was based on incomplete data. Because Dr. Siskin filed his
Liability Phase Report in November 2007, he did not (and could not)
take into account the last class that the City hired from the 2043
list in late January 2008. (See City Mem. in Opp. to Backpay Mot. for
Summ. J. at 3.) The City argues that Dr. Siskin's shortfall
calculation for the
2043 list was therefore inaccurate and should be ignored.*fn3
(Id. at 6-7.) The City further argues that a mere reference
to Dr. Siskin's shortfall calculation in the Disparate Impact
Liability Opinion does not mean that the court should use that
calculation; the City claims that the shortfall number was not
consequential to the finding of disparate impact liability and
therefore does not carry over to the monetary relief phase of the
case. (Id. at 4, 6-7, & n.7.)
The City next claims that Dr. Siskin's methodology is faulty in three ways: the so-called "denominator deception," "PPT-ploy," and "attrition disappearance." (Id. at 8-12.) The City's first critique is its claim of "denominator deception": According to the City, "Dr. Siskin creates an artificial construct of 'effective passers'-claimed to be those who achieved at least the score of the lowest-scoring person actually appointed." (Id. at 8.) According to Dr. Erath, this construct distorted Dr. Siskin's results in a manner overly favorable to minority applicants. (Erath Decl. & Report at ¶¶13-15.) The second problematic step in Dr. Siskin's methodology, according to the City, is the "PPT-ploy": "Dr. Siskin further employs the false presumption that everyone got 100 on the PPT . . . ." (City Mem. in Opp. to Backpay Mot. for Summ. J. at 10.) Finally, the City argues that Dr. Siskin engages in "attrition-disappearance": "there is a greater attrition rate among black and Hispanic candidates, as compared to white candidates . . . . Dr. Siskin seeks to hide this disparity by choosing to apply an attrition rate based upon all ethnicities." (Id. at 11.) As evidence for a greater attrition rate among minority candidates, the City cites Dr. Erath's observation that "the fraction who self-selected out of the process or were judged ineligible for reasons unrelated to the claims in this case was larger for those with higher (i.e., less advantageous) list numbers . . . ." (Erath Decl. & Report at n.2.) Each of these steps, according to the City, inflated Dr. Siskin's estimate of the number of shortfall hires.*fn4
The City also objects to Dr. Siskin's method of distributing shortfall candidates to academy classes. The City argues that shortfall candidates should not be distributed in the same proportion as all the hires because: first, minority candidates were more likely to be delayed in processing than white candidates (as mentioned in the discussion of the "attrition-disappearance" problem the City perceives in Dr. Siskin's methodology); second, minority candidates, on average, scored lower than white candidates on the PPT, which would factor into their ranking; and, third, performance on the PPT was correlated to performance on the written exams. (City Mem. in Opp. to Backpay Mot. for Summ. J. at 12-13.) Dr. Erath's Report proposed distributing the shortfall candidates to either the last classes hired from the relevant list or distributing them to the classes hired after the median hire date. (Erath Decl. & Report at 18-19.)*fn5
The court will first address the City's objections to Dr. Siskin's shortfall hire methodology; then, the court will address the issue of the January 2008 fire academy class second; the court will discuss the City's objection to Dr. Siskin's method of distributing shortfall hires to fire academy classes last. The City's objections to Dr. Siskin's methodology must be rejected for several reasons. The first is that the number of shortfall hires has been previously determined in the court's Disparate Impact Liability Opinion, and is thus law of the case. The law of the case doctrine "commands that when a court has ruled on an issue, that decision should generally be adhered to by that court in subsequent stages in the same case unless cogent and compelling reasons militate otherwise." Johnson v. Holder, 564 F.3d 95, 99 (2d Cir. 2009) (citations and quotation marks omitted). The City suggests that the Disparate Impact Liability Opinion had not resolved this issue. (City Mem. in Opp. to Backpay Mot. for Summ. J. at 4 n.7.) A careful reading of the court's Disparate Impact Liability Opinion, however, shows otherwise. The decision quoted Dr. Siskin's shortfall estimates extensively, as evidence of the practical importance of the statistical disparities in 7029 and 2043 passage rates. The court concluded that "absent these disparities [in pass rates and rank-ordered results on Exams 7029 and 2043] approximately 293 additional black and Hispanic candidates would have been appointed from the eligibility lists used from 2001 through 2008 . . . [I]t is clear that these disparities have a substantial practical significance." (D.I. Op. at 27 (emphasis added).) The court also noted that that "the City has essentially admitted the calculations performed by Plaintiffs' experts showing the disparities' practical significance." (Id. at 28.) The court's liability ruling "decided the issue" of the methodology of calculating and the resulting number of shortfall hires that the underlying disparities in pass rates produced. Moreover, to the extent that was unclear to any party, the court then relied extensively on that same number of shortfall candidates in its Initial Remedial Order. (See I. R. Order at 20-30.) The court and the parties have used these calculations, and relied on the methodology that underlies them, for years.
The City offers no cogent reason for reversing this determination. The City did not oppose this methodology in calculating a shortfall either in its Opposition for Motion for Summary Judgment or in its briefing filed in anticipation of the Initial Remedial Order. The prejudice that such a reversal would create to the United States, the Plaintiff-Intervenors (whose expert relies on Dr. Siskin's work), and the potential claimants waiting for some form of relief for injuries that reach back to 2001 would be substantial. The court will not reverse itself after the City has failed to utilize so many opportunities to object to the methodology of the United States' principal statistical expert and at a point where the prejudice that would result would be substantial.
Even if the court were inclined to reconsider the issue of the number of shortfall candidates, it would not reach a conclusion contrary to its original number. The City's arguments against Dr. Siskin's the methodology are the so-called "denominator deception," "PPT-ploy," and "attrition disappearance" discussed above. (City Mem. in Opp. to Backpay Mot. Summ. J. at 8-12.) The City's brief is parsimonious in its citations, declining to cite to Dr. Siskin's report, and so it is not entirely clear where the City believes Dr. Siskin commits these errors; however, it appears that the City's first two arguments relate only to Dr. Siskin's method of calculating the number of shortfall candidates who were not hired because the 2043 list was not exhausted. Here, the City evidently is confused by Dr. Siskin's report. Dr. Siskin's report included two different analyses-the "effective failure rate" and the "rankings test"-which were two different approaches for the same task: determining the shortfall created through an unexhausted 2043 list. (Compare Siskin Liability Report at 28-31 & Tables 7 & 9 with Siskin Liability Report at 32-35 & Tables 11 & 13.) The City's objections appear to be targeted toward Dr. Siskin's first mechanism, which Dr. Siskin described as "effective failure." Dr. Siskin's effective failure analysis attempted to determine what score a candidate would have needed to earn to have a chance at being hired if that candidate scored a 100 on the PPT; it produced a shortfall of about 70 black firefighters. (See Siskin Liability Report at 29-30 & Table 7.) Dr. Siskin's other form of analysis-the rankings test-used data about candidates who actually had been ranked on the 2043 list-i.e., it incorporated candidates' actual Written Exam 2043 and PPT scores, and, when combined with the shortfall of the candidates who truly failed the Exam, produced a shortfall of 72 black firefighters. (See Siskin Liability Report at 32-33 & Tables 5 & 11 (combined shortfall from nominal failure rate and "rankings test" pass rate is 72 black firefighters).) As the United States correctly notes, the court used the latter number-72-in its liability opinion findings on the shortfall in black firefighters from Exam 2043. (United States Reply Mem. in Support of Backpay Mot. for Summ. J. (Docket Entry #547) at 4.) Thus, the City's objections to the "effective failure" portion of Dr. Siskin's report are simply irrelevant: The court adopted the rankings test, and Dr. Siskin and the United States accordingly based their estimates of damages on the numbers generated by that test. (Siskin Relief Phase Report at 2.)
The City's third argument-the attrition disappearance claim-is presumably applicable to all of Dr. Siskin's non-hire shortfall calculations (i.e., his calculations for those who failed 7029, those who failed 2043, and those who passed 2043 and were never hired because of their disadvantageous ranking). The only basis that the City offers for its contention that black and Hispanic candidates suffered from a higher rate of attrition is Dr. Erath's statement that individuals with less advantageous rankings were less likely to report for processing and submit required documents. (See Erath Decl. & Report at 15 n.2.) However, as has already been discussed in some detail above, minority candidates were more likely to be ranked less advantageously than white candidates. See discussion of D. I. Op. in Part I supra. The City's evidence does not establish an attrition rate independent of ranking, i.e., independent of the Exams. Dr. Siskin's assumption-a uniform attrition rate across ethnic groups (see Siskin Liability Report at 16-17)-is the only mechanism apparent for avoiding the discriminatory effect of the invalid exams.
The court next considers whether to modify the number of shortfall hires to take into account the results of the January 2008 academy class. The court will not do so. The City's actions in this regard were governed by the discovery provisions of the Federal Rules of Civil Procedure. The City is certainly correct that if the last group of firefighters from the 2043 list were hired in January 2008, then it could not have informed the parties about that group of hires until after Dr. Siskin produced his November 2007 liability report. However, the City offers no reason why it did not fulfill its discovery obligations to Plaintiffs by informing them of additional hires once the hiring process was complete. See Fed. R. Civ. P. 26(e)(1)(A) ("a party who has made a disclosure . . . or who has responded to an interrogatory, request for production, or request for admission-must supplement or correct its disclosure or response . . . in a timely manner if the party learns that in some material respect the disclosure or response is incomplete or incorrect . . . .") Had the City done so, Dr. Siskin could have updated his report before it was filed as part of the United States' motion papers. Instead, the City waited for a year and a half to disclose the last set of data relating to hiring off the 2043 list. (United States Reply Mem. in Support of Backpay Mot. for Summ. J. at 2.) The City's failure to follow the Rules of Civil Procedure created the incomplete nature of Dr. Siskin's data, and the City has not explained why this information should not be excluded as the Rules of Civil Procedure suggest: "if a party fails to provide information or identify a witness as required by Rule 26(a) or (e), the party is not allowed to use that information or witness to supply evidence on a motion, at a hearing, or at trial, unless the failure was substantially justified or is harmless." Fed. R. Civ. P. 37(c)(1).
Ignoring for a moment the rules of discovery, the City's argument is still unavailing because, as discussed above, the shortfall hire number is the law of the case; the City has had many opportunities to urge the court to change its position and has not done so until now. One "cogent and compelling reason," Johnson v. Holder, 564 F.3d at 99, for a court to revisit its prior determination is the "availability of new evidence," United States. v. Quintieri, 306 F.3d 1217, 1230 (2d Cir. 2002). However, the January 2008 fire academy class was not "new" when the court was considering Plaintiffs' Motion for Summary Judgment, filed in March of 2009. This information was available to the City well before it served its opposition papers, and well before the court issued its Disparate Impact Liability Opinion. Thus, the City can not avail itself of this reason to depart from the court's previous rulings. Nor could the City move for reconsideration on this issue, as such a motion would be both untimely, Fed. Rule Civ. P. 59(e) (motions to amend a judgment should be brought within 28 days of that judgment), and not permitted for the purpose of introducing "new" evidence that was available to the movant before the court issued its opinion. Brown v. J.F.H. Mak Trucking, No. 95-cv-2118 (JS) (JO), 1999 WL 1057274, at *1 (E.D.N.Y. No. 8, 1999).
Finally, the court notes that incorporating the additional hires that the City belatedly disclosed but otherwise using Dr. Siskin's methodology would reduce the number of shortfall candidates by only five. (Siskin Relief Phase Rebuttal Report (Docket Entry # 538-5) at 9-11.) Thus, the City will suffer relatively little prejudice. Moreover, adhering to the court's previous finding on the number of shortfall candidates will not impact the rights of any individual claimant vis-a-vis another, nor will it impact the rights of any third parties.*fn6 In sum, the court sees no reason to revisit what it has already decided to consider information that the City could have presented to the parties and the court and did not.
The City's objections to Dr. Siskin's method of allocating shortfall candidates to different academy classes-i.e., Dr. Siskin's assumption that the 293 shortfall hires would have been hired to each class in the same proportion as the rest of the candidates-are predicated on two arguments. The first is that black and Hispanic candidates "failed" the hiring process due to attrition at a higher rate than did white candidates and the second is that black and Hispanic candidates scored lower on the PPT on average than did white candidates. (City Mem. in Opp. to Backpay Mot. for Summ. J. at 12-13.) The first argument has already been addressed in the context of discussing the City's "attrition disappearance" claim, and the court rejects it for the reason discussed.
The second assumes that the PPT would be a valid rank-order device if used alone, and so its results should be incorporated into the court's estimate of the gross loss of backpay. The City is certainly correct that the court has never ruled that the PPT was not job-related. (City Mem. in Opp. to Backpay Mot. for Summ. J. at 14.) However, the City misunderstands the requirements of Title VII. Once a prima facie disparate impact case has been made out, the employer has the chance to prove that an employment exam is job-related. See Gulino v. New York State Educ. Dep't, 460 F.3d 361, 385 (2d Cir. 2006). The City does not claim to have submitted any evidence from which the court could deduce that rank-order usage of the PPT, by itself, is job-related. Plaintiffs, however, have offered evidence that use of the PPT as a rank-order device is not job-related. (Jones & Hough Expert Report (Docket Entry # 264-4) at 58-60.) Therefore, the City's concession that the PPT produces an adverse impact does not help its position-because the City has not even attempted to offer evidence that using the PPT as a ranking mechanism would be justified by business necessity. Hence, the court will not assume that a disparity in PPT passage rates among different ethnic groups is legally justifiable. Nor will the court reject Dr. Siskin's report because of his use of a uniform PPT passage rate for all ethnic groups.
Finally, the court notes that the parties accuse each other in their briefs of not assuming parity among ethnic groups. (Compare United States Mem. in Support of Mot. for Backpay Summ. J. (Docket Entry # 536) at 15 with City Mem. in Opp. to Backpay Mot. for Summ. J. at 13 and United States United States Reply Mem. in Support of Backpay Mot. for Summ. J. at 6.) Most importantly, the City alleges that if the United States' candidate shortfall number and distribution of those candidates to fire academy classes were adopted, the City would have been required to treat black and Hispanic candidates better than white candidates and so engage in disparate treatment of white firefighter candidates. (City Mem. in Opp. of Mot. for Summ. J. at 12.) Firstly, the City's contention is a non sequitur-the United States is not asking for a remedial order that requires the City to follow Dr. Siskin's estimates and calculations as hiring policy; it is asking for a remedial order that provides monetary relief for victims of the invalid Written Exams and accepts certain estimates in doing so, because the City's discriminatory practices have placed the parties in the position of evaluating the likelihood of counterfactual events. Secondly, the City's contention is simply incorrect. As the United States notes, even once the shortfall hires of 72 black firefighters and 45 Hispanic firefighters are added to the minority firefighters hired from the 2043 list, there is still a disparate success rate among ethnic groups: black candidates would have a 11.84% success rate, Hispanics a 10.83% success rate, and white candidates a 15.20% success rate. (United States Reply Mem. in Support of Backpay Mot. for Summ. J. at 6.)
Dr. Siskin's next step in his analysis was to estimate the income each shortfall hire would have made. Using data that the City provided in discovery disclosing how much each firefighter from each academy class earned per year, Dr. Siskin was able to compute the average earnings of a firefighter from a particular academy class per year (assuming a firefighter stayed on the job for a full year each year). (Siskin Relief Phase Report at 6.) The City does not object to this step in Dr. Siskin's methodology, and Dr. Erath did not disagree with it in his expert report. (City Mem. in Opp. of Mot. for Summ. J. at 2; United States Backpay Motion Rule 56.1 Statement (Docket Entry # 544) ¶ 124.)
Dr. Siskin's third step was to discount the average earnings of a firefighter by the likelihood that a shortfall hire would have left the FDNY's employment or taken unpaid leave. (Siskin Relief Phase Report at 6-7.) Dr. Siskin did this by using data the City provided that showed the number of firefighters who had taken leave without pay or left the FDNY between 2001 and 2009. (Id. at 7 n.6.) Dr. Siskin created an average probability that a firefighter hired from a particular academy year would have left or been on leave in a given subsequent year; then he discounted the earnings an average firefighter made per year based on that probability. (Id. at 7.) Neither the City nor Dr. Erath ...