Full opinion text
MEMORANDUM NEWCOMER, District Judge. On July 18, 1983 this Court held that defendant United States Steel Corporation (USS) violated Title VII by engaging in race discrimination against the class of black applicants who unsuccessfully sought employment in the production and maintenance department (P & M) at the USS Fairless Hills Plant during the following two periods: July 11, 1972 through December 31, 1974, and January 1, 1978 through December 31, 1979 (the “class period”). Green v. United States Steel Corp., 570 F.Supp. 254 (E.D.Pa.1983) (liability opinion). At the liability stage, plaintiffs proved, through the use of statistical evidence, that USS failed to hire black applicants in proportion to their percentage in the applicant pool, and that this failure had a discriminatory impact on blacks. This Court found that USS, in its hiring procedures, relied on an amalgam of largely subjective criteria, and its interviewers were not trained so as to ensure even-handed application of the procedures. Id. at 260. Moreover, the minimum hiring criteria for P & M positions were so low that they did not eliminate many applicants. The liability opinion explicates in great detail this Court’s findings and the reasons therefore, and I incorporate it as if set forth herein. Only those portions essential to an understanding of this opinion will be repeated when necessary. Following the liability opinion, the Court directed the parties to meet, identify issues pertinent to relief, and develop any information necessary to calculate the damages to be awarded to the class. Order of January 12, 1984. The parties were able to agree on a basic framework, to compile a stipulated set of damages calculations (Ex. P-200), and to focus on the several remaining contested legal issues. Both sides fully briefed the following issues, now ripe for this Court’s disposition: 1. Whether applicants for summer positions in 1972 and 1973 are included in the class. 2. Whether the decertification notice sent in Dickerson precludes unsuccessful black applicants in 1972, 1973 and 1974 from recovering damages in Danley. 3. The proper USS employee reference group to use to compute back pay and attrition rates. 4. Which fringe benefits, if any, should be included in a back pay award. 5. The appropriate method for calculation of mitigation of damages (interim earnings). 6. Whether USS should be given a “credit” for those years in which they did not discriminate against, black applicants. 7. Whether plaintiffs are entitled to front pay, and if so, how it should be calculated. 8. Whether plaintiffs are entitled to prejudgment interest, and if so, how it should be computed. 9. Whether plaintiffs are entitled to any additional monetary compensation to account for the tax effect of lump sum payments. 10. Who should bear the costs relating to the distribution of damages (including the cost of notice to the class members). 11. Whether plaintiffs are entitled to injunctive relief and, if so, how it should be fashioned. The parties also presented testimony, documents and arguments to the Court at the damages trial held April 7-14, 1986. This Memorandum Opinion shall constitute the Court’s findings of fact and conclusions of law, as required by Fed.R.Civ.P. 52(a). FINDINGS OF-FACT I. Stipulated Framework for Assessment of Damages 1. A central purpose of Title VII is “to make persons whole from injuries suffered on account of unlawful employment discrimination.” Franks v. Bowman Transportation Co., 424 U.S. 747, 763, 96 S.Ct. 1251, 1264, 47 L.Ed.2d 444 (1976) (citing Albemarle Paper Co. v. Moody, 422 U.S. 405, 418, 95 S.Ct. 2362, 2372, 45 L.Ed.2d 280 (1975)). 2. Title VII gives district courts significant latitude to “order such affirmative action 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). 3. The goal at the damages stage is to award plaintiffs complete relief in order to make them whole, while at the same time not awarding plaintiffs damages which will punish defendant or give plaintiffs a windfall. 4. The assessment of damages has been facilitated by numerous factual and legal agreements reached by the parties before trial. The Court finds these agreements, noted and summarized below, to afford a fair and reasonable approach to the calculation of damages owed the plaintiff class. 5. USS conceded that back pay was an appropriate element of damages due the plaintiff class, and that this component “should be based upon the average annual wages earned by persons who were actually hired by USS in each of the liability periods.” (Ex. P-200, Section IV at 11). 6. The Court finds it reasonable to calculate back pay in this fashion because, by definition, the plaintiff class was not hired by USS and therefore does not have an employment history at USS. 7. Moreover, the liability period spans almost ten years, the number of class members (approximately 13,000) greatly exceeds the jobs available even in the absence of discrimination (386). The subjective nature of the hiring criteria as employed by untrained interviewers, formed the factual basis for the liability opinion. 570 F.Supp. at 260, 269. Hence it is not practical to attempt to identify which members of the plaintiff class would have been hired but for the defendant’s race discrimination. 8. Because the calculation of damages on an individual basis is not a feasible way to proceed, the parties have stipulated that damages should be calculated and distributed on a class-wide basis. (Ex. P-200 at 1). This approach has been adopted by other courts. See, e.g., Hameed v. International Ass’n of Bridge Workers, Local 396, 637 F.2d 506 (8th Cir.1980); Stewart v. General Motors Corp., 542 F.2d 445, 452 (7th Cir.1976), cert. denied, 433 U.S. 919, 97 S.Ct. 2995, 53 L.Ed.2d 1105 (1977); Pettway v. American Cast Iron Pipe Co., 494 F.2d 211, 261 (5th Cir.1974) (“[W]hen the class size or ambiguity of promotion or hiring practices or the multiple effects of discriminatory practices or the illegal practices continued over an extended period of time calls forth the quagmire of hypothetical judgment ... a class-wide approach to the measure of back pay is necessitated”). 9. The basic approach used to measure the plaintiffs’ damages has been to ascertain how many class members would have been hired in each liability period in the absence of discrimination (the Black Hiring Shortfall); compute the value of wages and fringe benefits those persons would have received each year from USS; and determine the amounts of wages and fringe benefits those persons could be expected to have earned each year at other jobs after not being hired by USS. The annual excess of USS wages and fringe benefits over outside wages and fringe benefits is the net damages lost that year. 10. There are five liability periods; July-December 1972, and all of 1973, 1974, 1978 and 1979. 570 F.Supp. at 257, 263. Damages are calculated separately for each group of blacks who should have been hired during each liability period (i.e., for each of the five “year class” cohorts), and for each such cohort to make damages calculations on a yearly basis, commencing with the year they should have been hired through the end of 1985. (Ex. P-200). 11. There were hiring shortfalls for permanent jobs in all five liability periods. Summer hiring shortfalls occurred in 1972 and 1973. 12. The annual shortfall figures were obtained by subtracting the actual number of blacks hired by USS as P & M employees in each liability period from the number of blacks who should have been hired in that liability period, had black applicants been hired in proportion to their representation in the applicant pool. The latter number was derived by multiplying the total number of USS hires in a liability period by the percentage of blacks then in the applicant pool. 13. This formula is in accord with this Court’s liability ruling that, absent discrimination, blacks would have been hired by USS in proportion to their representation in the applicant pool. 14. The aggregate USS Black Hiring Shortfall, including both permanent and summer jobs, is 386 persons. (Ex. P-200, Section II). a. The Black Hiring Shortfall for permanent jobs in each liability period is as follows: 1972 _ 19 people 1973 — 101 people 1974 — 82 people 1978 — 74 people 1979 — 52 people TOTAL — 328 people b. The Black Hiring Shortfall for summer jobs in the pertinent liability periods is as follows: 1972 — 6 people 1973 — 52 people TOTAL — 58 people 15. The Black Hiring Shortfall for the summers of 1972 and 1973 are “borrowed” from the figures for permanent employees for those years because the summer applications for 1972 and 1973 were destroyed. 16. The Black Hiring Shortfall is adjusted in each year subsequent to the year that a cohort class would have been hired, to reflect actual attrition by USS employees in the P & M department. This ensures that damages are not paid for employees who would no longer be employed by USS in the normal course of events (i.e. even in the absence of discrimination). (Ex. P-200, Section III). 17. The attrition rates for P & M employees at USS were calculated and set forth separately in the Stipulated Damages Figures for white USS employees, black USS employees and all USS employees hired during the liability periods. (Ex. P-200, Section III). II. Whether Summer Applicants for 1972 and 1973 Are Included in the Class 18. Fifty-eight people comprise the Black Hiring Shortfall for the summers of 1972 and 1973. The class certified by Order dated August 16, 1982 consisted of: All black persons who unsuccessfully sought employment at the Fairless Hills Plant of United States Steel Corporation between July (11), 1972 and the present. 570 F.Supp. at 256. 19. At the liability trial, the plaintiff class refined its liability periods of proof to exclude 1975-1977, 1980 and 1981. See note 1, supra. 20. At no time were the 1972 and 1973 summer applicants excluded, and in fact several references to summer applicants, in combination, support the inclusion of 1972 and 1973 summer applicants. See Findings of Fact Nos. 16, 23, 83,123 and 128 (“Moreover, [Talley’s] figures do not include applications submitted by persons seeking summer jobs, although they are members of the class.”); Conclusions of Law Nos. 4-6 (summer hires included for the liability years; this Court explicitly excluded any portions of the certified class and unsuccessful theories in the liability opinion’s conclusions of law). III. Decertification Notice in Dickerson and its Effect, if any, on the 1972-197.4 Members of the Danley Class 21. The Court incorporates by reference as if set forth herein the Memorandum and Order of March 31, 1981, where this Court addressed this issue and found that, based on the timely filing of the Danley action after the Dickerson notice became effective, the 1972, 1973 and 1974 unsuccessful applicants are included in the Danley class. IV. The Proper USS Employee Reference Group For Computation of Back Pay and Attrition Rates 22. There is an unexplained disparity between USS wages paid to white employees and black employees during the liability years. (Ex. P-200, Section IV). 23. Plaintiffs advocate using white USS employees as the reference group for measuring back pay, on the grounds that: white applicants took the place of those black applicants who were not hired by USS; and if white USS employees do not provide the reference group, then the class members will be penalized twice — once for not being hired, and again in computation of damages. 24. While plaintiffs’ position has some initial appeal, upon closer inspection their position is flawed. There was no evidence from which I could reasonably conclude that the acknowledged wage disparity resulted from discrimination against black employees, once hired at the Fairless Hills plant. 25. Defendant takes the position that the black USS employees are the proper reference group, as the group “having employment characteristics most closely resembling those of plaintiffs and who were actually hired at that facility.” Defendant's trial brief at 14. 26. This argument is inherently offensive, as plaintiffs point out, because the suggestion is that black employees perform like one another simply because they are black. The racial composition of a group of workers is not a legitimate “employment characteristic” upon which this Court may rely. 27. The Third Circuit has not yet been faced with the situation presented here, nor even with an analogous problem. The cases relied upon by both parties (from other circuits) do not provide much guidance in this area and thus I am faced with a novel issue. 28. With the exception of one case cited, plaintiffs cite cases which are inapposite because the cases involve individual calculations and/or discrimination in post-hiring employment practices (training, promotion, etc.). 29. These cases are not persuasive precedent because those courts usually had prior work histories with the employment characteristics for the class members involved. Also in measuring damages for one individual, a personal history could be obtained to find the closest match with another employee. 30. For example, plaintiffs cite to Pettway v. American Cast Iron Pipe Co., 494 F.2d 211 (5th Cir.1974), a class action suit involving discrimination by a cast iron and pipe manufacturer in its promotion practices. Id. at 217 & n. 5. There the court instructed the trial court, on remand, to assess back pay damages on a class-wide basis. In response to a suggestion by the Equal Employment Opportunity Commission (EEOC), the Pettway court recommended use of a class-wide formula “ ‘based upon a class of whites which would be comparable to the members of the effective class but for the discrimination,’ ” id. at 263 (emphasis supplied), because this method takes into account “actual advancement of a comparable group not discriminated against.” Id. (emphasis supplied). Here there is no evidence of racial discrimination in post-hiring employment practices. 31. Individual advancement at USS is highly dependent upon such employment factors as overtime and individual incentive. Thus, in the absence of any evidence of racial discrimination in USS post-hiring employment practices, I cannot find it equitable or a good “match” to use white USS employees as the reference group. 32. Plaintiffs also rely on the Seventh Circuit case of Stewart v. General Motors Corp., 542 F.2d 445 (7th Cir.1976), cert. denied 433 U.S. 919, 97 S.Ct. 2995, 53 L.Ed.2d 1105 (1977), for the proposition that in a Title VII class discrimination suit alleging racial discrimination in hiring and promotion at one of defendant’s plants, white employees were the proper reference group for back pay. 33. The Stewart court, in its opening paragraph, describes the case as one “concernpng] claims of racial discrimination in hiring and promotion against General Motors Corporation at its Broadview, Illinois Parts Distribution Center” by a class of “black people who have been employed at this facility in hourly rated positions since December 21, 1983.” Id. at 449 (emphasis supplied). However, a close reading of Stewart reveals that the suit deals only with discrimination in two types of promotions, and not at all with hiring practices. Consequently, when the court recommends use of white employees as the reference group for a class-wide remedy, this suggestion encompasses a need to remedy racial discrimination against class members already hired by the defendant. See 542 F.2d at 453. 34. One case has advocated the use of whites as the reference group in a suit claiming racial discrimination in hiring, where the damages were calculated on a class-wide basis. Hameed v. International Ass’n of Bridge Workers, Local 396, 637 F.2d 506 (8th Cir.1980). 35. In Hameed, blacks were denied entry into ironworkers apprenticeship programs from 1965-1973. Liability in Ha-meed was based on a disparate impact theory and the court calculated damages using the Black Hiring Shortfall model. In explaining to the district court its task for back pay computation, the Hameed court described back wages using white apprentices admitted to the program. Id. at 520-521. Hameed cites, inter alia, to Stewart, supra, for guidance in its formulation of class-wide apportionment of damages. Nowhere does the Court in Hameed explain why whites are chosen as the reference group. At best, it appears that the court did not reflect upon the issue of a reference group at all but rather, using Stewart as a model, assumed that white employees provided the correct reference group. 36. In the beginning of its back pay discussion the Hameed court set out the three issues of concern. Id. at 519. None of the three relates to which reference group is appropriate. Instead the court addresses the issues of class-wide versus individualized calculations, determination of the Black Hiring Shortfall, and a method to identify class members entitled to relief. Id. 37. Individual vicissitudes of employment earnings are too great to award the plaintiff class the highest back pay award available for workers in their putative division. I therefore find that the most equitable choice of employees for back pay (and as a corollary matter, attrition rates) is all USS P & M employees for the class periods (Ex. P.200, Section IV). I recognize, as I must, that taking the whole universe of P & M employees remains, at best, an imperfect fit, but it is the best choice in this case. As previous courts have frankly stated in cases involving a class-wide determination of back pay “[a]ny method is simply a process of conjectures.” Stewart, supra, 542 F.2d at 453 (citing Pettway, supra, 494 F.2d at 261). 38. The USS wages lost by the class members and calculated by reference to the Black Hiring Shortfall (386 persons) is determined as follows: On a yearly basis the aggregate wages USS paid to all employees actually hired as P & M laborers in each liability period is divided by the number of P & M laborers working for USS in the given year. This yields average annual per capital wages at USS for each year within each liability period. These average amounts are then adjusted for attrition (using the attrition rates of all USS P & M employees at Fairless Hills), to reflect what the Black Hiring Shortfall (386 people) would have earned while working at USS through December 1985 (Ex. P-200, Section IV). 39. The average earnings are calculated for each entering class year, or cohort, on a yearly basis, by multiplying the Black Hiring Shortfall for each cohort in a given year by the average yearly earnings per capita. Each year the Black Hiring Shortfall is reduced to reflect attrition. 40. Using the tables from the Stipulated Damages Figures (Ex. P-200, Section IV) with all USS employees as the reference group, yields USS wages for the Black Hiring Shortfall in the following amounts: a. Permanent Jobs Year of Hire Average Earnings for All USS Employees 1972 $ 1,470,714 1973 7,341,084 1974 5,627,332 1978 4,153,546 1979 1,414,712 TOTAL $20,007,388 Summer Jobs Year of Hire Average Earnings for All USS Employees 1972 $ 15,192 1973 137,072 TOTAL $ 152,264 ÍALL TOTAL $20,159,652 V. Fringe Benefits 41. In addition to “straight salary” as a component of back pay, the plaintiff class is entitled to an award of fringe benefits. See, e.g., Rasimas v. Michigan Dept. of Mental Health, 714 F.2d 614, 626 (6th Cir.1983), cert. denied, 466 U.S. 950, 104 S.Ct. 2151, 80 L.Ed.2d 537 (1984); United States v. Lee Way Motor Freight, Inc., 625 F.2d 918, 945 (10th Cir.1979); Pettway, supra, 494 F.2d at 263; Rosen v. Public Service Electric & Gas Co., 477 F.2d 90, 96 (3d Cir.1973); Shaffield v. Northrop Worldwide Aircraft Services, Inc., 373 F.Supp. 937, 945 (M.D.Ala.1974). The dual purpose of this award is to make plaintiffs whole and to deter defendant from future wrongdoing. 42. During the relevant time period, USS fringe benefits for laborers were better than those generally available for similar positions in the outside labor market. 43. In the earlier years, the cost of USS fringe benefits exceeded 50% of the hourly USS wages paid (Ex. P-200, at 104-07). 44. USS actively negotiates with the Steelworkers’ Union for the salary package, which includes fringe benefits as a significant amount of each worker’s total salary. 45. USS conceded that, as a general principle, the plaintiff class is entitled to some fringe benefits. However, defendant maintains that plaintiffs have not proven any loss and so have not sustained any recognizable damages for fringe benefits. In other words, defendant claims that each member of the plaintiff class must prove an actual loss prior to this Court awarding any fringe benefits. 46. I find that the fringe benefits were of great value to the plaintiff class. Accordingly, the question is how to calculate the allowable fringe benefits, and not whether to grant them at all. 47. A reasonable method for computation of fringe benefits for the class is to use the fringe benefits received by those USS P & M employees who were hired during the liability class years. 48. It is reasonable to conclude that the plaintiffs, had they been hired, would have received at least as much benefit from USS fringe benefits as actual USS employees received. 49. USS acknowledged that it would be impossible to ascertain the dollar value of USS fringe benefits received by each person who actually was hired by USS during each liability period. 50. USS is a self-insurer, and calculates its cost on an actual loss basis. At plaintiffs’ request, USS provided the average cost to USS in each liability year for several fringe benefits. Notwithstanding these computations USS maintained that cost was not the proper measure of the value to plaintiffs of the fringe benefits. (Ex. P-200, Section V). 51. I find that the use of USS cost, rather than computation of loss for each member of the plaintiff class, is a reasonable method for calculation of fringe benefits to award plaintiffs under the circumstances of this case, for the reasons set forth below. 52. Plaintiffs are clearly entitled to some amount of some fringe benefits. USS cannot simply claim that the cost method is inaccurate without offering an alternative affirmative method of calculation. In this situation all of the relevant factual damages information resided with deféndant, and was most easily usable or computed, by USS employees, if by anyone at all. While I remain aware that plaintiffs bear the burden of demonstrating their damages, when it is clear, as here, that plaintiffs are entitled to a fringe benefit award, I find it incumbent upon defendant to do more than simply say it does not agree with plaintiffs’ computation efforts. 53. Second, I find that average cost for a self-insuring company like USS is a reasonable method of computation, and does not excessively overestimate the damages. 54. Last, to the extent that the average cost amounts are excessive, USS as the wrongdoer should bear the loss. See, e.g., Craig v. Y & Y Snacks, Inc., 721 F.2d 77, 83 (3d Cir.1983). 55. USS presented the testimony of George Hensarling, USS Director of Employee Benefits for the past eight years, and an employee of USS for 38 years. Mr. Hensarling negotiates the contract with the Steelworkers’ Union, and is responsible for administering all of the USS fringe benefits. 56. Mr. Hensarling testified that following USS’s most recent contract negotiations with the union (March 1983), the base wage rate agreed upon for non-incentive jobs was $12/hour, with $ll/hour for unworked hours — fringe benefits (such as insurance, pensions, etc.) and vacation, holidays and funeral leave. 57. During the middle of the damages trial a dispute arose as to the interpretation of the stipulated damages figures for fringe benefits. 58. Specifically, defendant argued that the fringe benefit costs in the stipulated damages were overinclusive, while plaintiffs believed these numbers represented the average fringe benefit costs to USS for the plaintiff class. For the reasons developed below, I find plaintiffs’ position prevails. 59. During the damages discovery period plaintiffs asked USS to state the average per capita cost or value, for each year from 1972 through 1985 for each of the pertinent USS fringe benefits provided to USS P & M employees hired during the liability periods. (Exs. P-201, No. 5; P-203; P-204). This request sought the average yearly cost for the class members, had they been hired, beginning with the year of application and continuing through 1985. 60. On the same day as plaintiffs’ request, defendant agreed to calculate the following: For each of the years for which liability has been found (July 11 through the end of 1972, 1973, 1974, 1978 and 1979) USS will identify by year class the total number of persons hired and for such hires the dollars paid as earnings, fringe or benefit payments for all subsequent years through December 1983 [later updated through December 1985]. We will determine the average dollars paid per individual per year class for each year subsequent to hire. This figure will be used as a base point for calculating back pay for each year class, from which will be subtracted the mitigation factor — i.e. average expectable earnings or benefits had the individual not worked at USS. Letter of January 10, 1984 from USS counsel to plaintiffs’ counsel. 61. Defense counsel added the caveat that the above calculations would be made to the extent possible using USS’s data, and of course, without waiving USS’s right to litigate the issue of plaintiffs’ entitlement to fringe benefits and the use of cost as a proper measurement of damages. 62. What is clear even from the initial contacts during discovery is that the parties agreed to the accuracy of the calculations, and agreed to use these cost calculations as the measurement of fringe benefits for the plaintiff class, should the Court find plaintiffs were entitled to fringe benefits. Further they agreed that cost to USS was a reasonable method of measuring the value to plaintiffs in the absence of another reasonable method. 63. Plaintiffs’ supplemental discovery request of October 31, 1985 reiterated their request for the cost of average fringe benefits for class members. See Ex. P-204, for the definition of fringe benefits: “Fringe benefits” means all compensation, and benefits other than “earnings,” provided to, for or on behalf of P & M hires, including F.I.C.A. contributions made by USS, unemployment compensation, supplemental unemployment benefits, pension, workers compensation and insurance (health, life, dental, S & A, and vision). 64. Using this definition, plaintiffs’ request for admission No. 5 stated: Thomas P. Preston, Esq.’s letter of September 27, 1984, at page 3 (a copy of which is attached and incorporated) correctly states the total dollar cost to USS, per employee, by year, of each of the benefits listed therein, specifically “F.I. C.A.,” Unemployment Compensation, Supplemental Unemployment Benefits, Pension Benefits, Workman’s Compensation and Insurance (including Health, Life, Dental, S & A, and union). Defendant admitted Request No. 5. 65. At this point, at worst, plaintiffs’ request for admission was inartful and a bit ambiguous, because plaintiffs ask for the “total dollar cost to USS, per employee,” without adding that the “employee” referred to is of course, one who would be a P & M hire, the counterpart to the plaintiff class, since those were the figures the parties had agreed to use for purposes of calculating cost. 66. USS gave plaintiffs the “total cost to USS per employee,” and thereafter, at successive times, always confirmed the accuracy of the calculations USS provided to plaintiffs up to and including the first day of the damages trial. (See Exs. P-200, P-207, P-208). 67. At no time before trial did USS challenge the appropriateness of using the costs it had supplied, for calculating the cost of fringe benefits lost by the plaintiff class, should this Court determine cost was a reasonable method of fringe benefit calculation. (See Exs. P-208, P-209). 68. The annual cost to USS of each pertinent fringe benefit is set forth in the stipulated damages figures. (Ex. P-200, Section V). 69. On the third day of the damages trial, for the first time USS (through Mr. Hensarling’s testimony) clearly revealed that its fringe benefit calculations not only included the plaintiff class, but also included any USS employees who are retired or who are on layoff, but who are still covered by USS benefits. Thus, according to USS, the calculations in the stipulated damages figures in Section V were grossly overinclusive. 70. Mr. Hensarling proceeded to testify to adjusted figures, using the stipulated damages as his base (Ex. P-200), to reflect the percentage of overestimate he believed would be a necessary reduction in order to reflect only P & M hires in a given year. 71. I reject Mr. Hensarling's percentage reductions for the fringe benefits of insurance, pension, and supplemental unemployment benefits (SUB), for several reasons. 72. As Mr. Hensarling admitted, USS did not disclose the substance of this portion of Mr. Hensarling’s testimony in the sworn interrogatory responses it served on plaintiffs’ counsel shortly before trial (identifying trial witnesses and providing requested summaries of their expected testimony). See Ex. P-209, No. 2(a). 73. Mr. Hensarling was not involved in the case until approximately one week before the damages trial. Up to that point, USS was relying on Section V of the stipulated damages figures as accurate reflections of how to calculate any fringe benefits awarded to plaintiffs, in the event the Court decided that average cost to USS was a reasonable method of calculating fringe benefits. 74. This Court finds it difficult to believe that USS maintained all along that the fringe benefit cost calculations represented cost to USS for employees other than the counterparts to the plaintiff class. 75. USS did all of the calculations, without any representative from plaintiffs reviewing the raw data. If, at some point during discovery, USS determined (as Mr. Hensarling later testified) that it is impossible to extract P & M hires from USS total cost for insurance, then it was incumbent upon defendant to notify plaintiffs of this limitation in the data source. 76. It strains credibility to argue, as USS does, that all along, it knew and that it so communicated to plaintiffs, that the insurance calculations were overinclusive. USS knew that plaintiffs were relying upon its accurate portrayal of fringe benefit cost to USS for P & M workers hired during the liability years, as the matching counterparts to the class. 77. During the damages trial USS did not offer the Court any specific numbers to substitute for those in the stipulated damages figures. The first time any revised calculations were provided was as attachment to defendant’s post-trial pleadings. These revised tables are not in evidence. 78. Nonetheless, even assuming that USS’s position is theoretically more accurate (thus putting to the side the misrepresentations to plaintiffs), USS did not present any credible evidence on which to base USS’s suggested reductions of the insurance benefit calculations in the stipulated damages. 79. Mr. Hensarling conceded that his percentage reductions were “guesstimates,” not based on more than his general sense of the people in the plaintiff class. Based on this speculation, Mr. Hensarling then offered percentage reductions of the insurance calculations, to exclude, inter alia, retirees and USS employees on layoff. 80. Mr. Hensarling also freely admitted that there were no “hard numbers” one could use to isolate insurance costs to USS for the plaintiffs’ counterparts during the liability years, because the method of insurance claims handling for USS is company-wide. In other words, all USS’s divisional insurance claims are lumped together. 81. Finally, to the extent the stipulated damages figures overestimate the fringe benefits due plaintiffs, the equities militate in favor of allowing these calculations to stand. As stated above, USS, not plaintiffs, were in possession of any and all data for the possible fringe benefit calculations. If these calculations would not isolate the plaintiff class, defendant had a duty to so inform plaintiffs. 82. I find that the most reasonable construction of plaintiffs’ discovery requests and defendant’s responses is that plaintiffs sought fringe benefits costs for the plaintiff class counterparts, and that plaintiffs justifiably and reasonably relied on defendant’s responses to provide calculations for that group alone. A. Insurance Benefits 83. The benefits provided by USS to its P & M employees under the insurance program consist of life insurance, sickness and accident benefits (S & A), hospital coverage, physicians’ services, major medical coverage, and dental and vision protection. (Exs. P-215 through P-221). Although USS employs an insurance carrier to process claims, USS is self-insured for its insurance program and thus is responsible for paying each claim. Mr. Hensarling described this as an “actual cost” system. 84. Mr. Hensarling testified that from the employees’ standpoint, the USS insurance program is very good. It provides “first dollar coverage;” that is, the employee is not required to pay anything — no deductible and no co-insurance. 85. The level of sickness and accident coverage and the amount of life insurance provided to employees at USS are tied to the employee’s wages, which is reviewed on an annual basis. 86. The USS sickness and accident program provides a source of income for employees who, as a result of illness or injury, are unable to work. 87. USS’s hospitalization program pays employees or dependents for inpatient or outpatient treatment, including maternity benefits. 88. USS’s physicians’ services and major medical coverages pay for other medical expenses of employees. 89. The USS vision program pays for eye exams and the cost of glasses. 90. USS’s dental benefits pay dental expenses of employees and their dependents. 91. The amounts set forth in the Stipulated Damages Figures are a fair and reasonable measure of the value of USS’s insurance benefits to its employees and, in turn, to the plaintiffs. (Ex. P-200 at 39-51). 92. USS tried to limit or eliminate the use of these figures with Mr. Hensarling’s testimony that each individual employee has hurdles to overcome (e.g., minimum 520 hrs. of work before covered, review by USS of a claim before it is paid, etc.) 93. I find that any of these concerns do not undermine the validity of using an average annual cost to USS for its employees, because on the whole, the average cost or payout by USS only reflects those people who have overcome those hurdles, and this Court has every reason to believe that the plaintiff class, on the average, would be similar to their counterparts who were hired by USS. 94. Insurance rates in the Stipulated Damages were calculated by multiplying the Black Hiring Shortfall for each year class (adjusted downward annually to account for attrition) by the average yearly insurance per capita (the cost to USS of providing insurance on a year-by-year basis for each relevant employee). This yields a total average yearly insurance amount per cohort. 95. For the reasons articulated as to straight salary portions of back pay, supra, I find that the proper reference group for insurance rates (and the other fringe benefits) is all USS employees. 96. Using the applicable charts in the Stipulated Damages (Ex. P-200, Section V) yields insurance benefits for the Black Hiring Shortfall in the following amounts: a. Permanent Jobs Year of Hire Total Aggregate Insurance 1972 $ 236,358 1973 1,310,298 1974 1,153,578 1978 1,201,005 1979 613,325 ALL YEARS TOTAL - $4,514,564 b. Summer Jobs Year of Hire Total Aggregate Insurance 1972 $ 4,406 1973 41,122 ALL YEARS TOTAL — $ 45,528 (For the summer jobs, USS provided percentages of the overall yearly cost which may be fairly attributed to summer jobs.) B. Pension Benefits 97. USS completely funds its pension plan, using actuarial estimates. Under the pension program, a USS employee who has ten years of continuous service receives a vested pension. 98. The type and amount of pension is determined based on such factors, inter alia, as age, continuous services at retirement and an employee’s earnings history. 99. When an employee is on layoff, he or she continues to accrue pension rights for a period up to two years. In addition, USS employees on layoff have recall rights for five years. Thus an employee who returns to active employment at USS within the five-year period maintains protected pension years accrued to date. 100. When an employee returns to active employment he or she receives another two year period for accrual of pension service. Thus, in order to receive a vested pension benefit, an employee need not actually work all ten years. For example: a member of the plaintiff class, if hired, could have worked for 4 years, then been on layoff for 2 years, worked for 2 years and then been placed on layoff again for 2 years. At the end of this ten year period, that individual would have a vested pension interest, although he or she worked for only 6 of the 10 years required for vesting. 101. The pension plan is funded on a group basis, not tied to an individual USS employee. Nonetheless, each year USS is obligated to make payments to the pension plan on behalf of every USS employee. The USS actuaries include in their annual calculation a factor for employee attrition. 102. As with the insurance calculations, Mr. Hensarling, at trial, testified to reductions in calculations for the plaintiff class pension benefits in the Stipulated Damages. 103. For reasons articulated above in the insurance section, I decline to adopt these eleventh-hour revisions. 104. In addition to the eleventh-hour nature of the calculations, Mr. Hensarling did not testify to data upon which I could rely. 105. The first time tables reflecting Mr. Hensarling’s revisions were presented to the Court was in defendant’s post-trial papers. These tables are not in evidence. 106. Moreover, this case from its inception has involved the use of complex and controlled statistical evidence. Mr. Hensarling admitted that his methods fell far short of this attempted precision. 107. For example, Mr. Hensarling testified that he “looked at” a recent liability year and found that approximately 38% of the funding went to employees for past service and 62% for current service. Based on this approximation (for which no documentation was supplied or even referred to) Mr. Hensarling decided to adjust the pension figures downward by 30%, to eliminate retirees who were not part of the plaintiff class. 108. Mr. Hensarling also conceded that although a USS employee’s pension benefit does not vest until ten years of accumulated service, in the earlier years the amounts contributed by USS are designed to cover collection by USS employees in future years. 109. Mr. Hensarling testified that USS sometimes permits individuals to receive immediate lump sum payments when they leave USS in lieu of deferred payments under the plan. 110. Many persons USS hired for P & M entry-level jobs during the liability periods now have vested interests in the USS pension. 111. Of the 328 members of the plaintiff class who should have been hired for permanent jobs, at least 70 would still be working at USS, and others on layoff would obtain future vested interests in the pension plan (Ex. P-200 at 4-10). 112. Mr. Braithwaite, USS Director of Corporate Employment (whose responsibilities include, inter alia, development of employment policies and procedures) testified that of the 1,486 P & M employees at Fairless Works currently on layoff, only about 100 of those have been laid off longer than two years. 113. Mr. Braithwaite also testified that all of the 1,486 P & M employees now on layoff may be recalled from layoff by 1990. 114. The amounts in the Stipulated Damages Figures are a fair measure of the per capita value to the plaintiffs of USS’s pension benefits. (Ex. P-200, Section V at 26-38). 115. As with insurance benefits, the cost of the pension benefit has been calculated by year class with the Black Hiring Shortfall in each subsequent year reduced to reflect attrition. This adjusted shortfall is then multiplied by the average yearly cost per capital, for a total average cost in that year. When totalled, this yields the aggregate average benefit for that year class. 116. Using all USS employees as the reference group for pension benefits yields the following pension benefits for the plaintiff class, using the Stipulated Damages Figures (Ex. P-200 at 26-38). Permanent Jobs Year of Hire Total Aggregate Pension 1972 $ 160,150 1973 875,392 1974 807,028 1978 871,174 1979 449,889 — $3,163,633 ALL YEARS TOTAL Social Security Benefits C. 117. John Re, a USS employee for 30 years, and currently a Manager in Tax Accounting, is responsible for all payroll tax withholding (federal, state and local income taxes, social security taxes, and federal and state unemployment taxes). He testified to the nature of the social security program as it affects USS and its employees. 118. Social Security tax is deducted from each USS employee’s pay. The employer pays an equal amount, and together the tax is paid to the U.S. Treasury on a periodic basis. 119. These amounts fund social security benefits. The payments in any given year fund social security benefits for current recipients. Thus specific funds are not designated to the currently employed individuals account, nor even segregated as to USS or another company. 120. In order to receive social security retirement benefits, an employee, at 62 years old, must have to his or her credit forty or more quarters of qualifying employment. 121. An individual’s earnings over the course of his or her work history determine the payments to which he or she is entitled under social security. Accordingly, the higher the earnings, the greater the benefits, although the rate of increase is not precisely proportional to the increased earnings. 122. Had the plaintiffs been hired, the quarters they worked for USS would be counted toward their social security coverage. Class members who were not hired by USS and did not find work elsewhere would receive no coverage toward the requisite forty quarters, because they would not be considered sufficiently attached to the labor market. 123. Had plaintiffs been hired, USS would have made social security tax payments on their wages. 124. Social security payments made by USS, while not tied to a particular employee, nonetheless confer a benefit on a USS employee which the plaintiffs would have received had they been hired. 125. Further, the USS wages — markedly higher than other laborer positions in the Philadelphia area — would have benefited plaintiffs by increasing their overall earnings. 126. At least two other courts have recognized that social security benefits are cognizable components of back pay recovery in discrimination cases. See Blim v. Western Electric Co., 731 F.2d 1473, 1480 (10th Cir.), cert. denied, AT&T Technologies, Inc. v. Blim, 469 U.S. 874, 105 S.Ct. 233, 83 L.Ed.2d 161 (1984) (social security payments by employer included in court’s back pay award in age discrimination case); EEOC v. Pacific Press Publishing Association, 29 Empl.Prac.Dec. (CCH) ¶ 32,907 at 26,297 (N.D.Cal.1982) (Title VII back pay award included social security tax payments, “on the basis that all monies paid are wages paid ... or untaxed fringe benefits, as the case may be.”). 127. Mr. Binkley, Staff Supervisor of Employment and Benefits at the USS Fair-less Hills facility, testified that the payment of employer social security contributions for a person who was improperly denied employment would be in keeping with the “make whole” concept. 128. Recognizing the inherent imperfection associated with the estimation of damages, yet keeping in mind the value of the social security payments to plaintiffs, I find that the amounts in the Stipulated Damages Figures áre a reasonable and fair measure of the social security benefits to which plaintiffs are entitled. 129. Using all employees as the reference group, the Stipulated Damages Figures yields the following social security benefits for the Black Hiring Shortfall: a. Permanent Jobs Year of Hire Social Security Benefits 1972 $ 132,667 1973 734,264 1974 638,680 1978 675,001 1979 344,899 $2,525,511 ALL YEARS TOTAL b. Summer Jobs Year of Hire Social Security Benefits 1972 $ 790 1973 8,019 ALL YEARS TOTAL — $ 8,809 D. Workmen’s Compensation Benefits 130. The purpose of the Pennsylvania Workmen’s Compensation Act is to compensate for wages lost by employees who are injured during the course of their employment. As a wage replacement, it has value to plaintiffs. 131. USS is self-insured as to workmen’s compensation, making payments to employees based on actual injuries. 132. Mr. Billy Helm, a USS employee who is Manager of the Workmen’s Compensation and Casualty, and who has 20 years experience in workmen’s compensation matters for USS, testified that USS annually places in reserve a specific dollar amount for estimated future liability. The cost to USS per employee is maintained on an annual basis in order to predict the future cost to USS. 133. The Pennsylvania Workmen’s Compensation Act protects P & M laborer employees at Fairless Works from their first day of employment. 134. Mr. Helm testified that because of USS’s high hourly wage rates, employees at USS who are injured in the course of employment are entitled to the statutory maximum amount of compensation set by the Commonwealth of Pennsylvania. 135. Workmen’s Compensation payments continue for as long as the disability continues. 136. Some employees hired by USS in each of the liability p'eriods have received Workmen’s Compensation payments. (Ex. P-209, No. 5(d)). Had plaintiffs been hired by USS, some of them would have received Workmen’s Compensation payments. (Ex. P-209, Nos. 2 and 5). 137. The amounts in the Stipulated Damages Figures reflect, on an annual basis, the average per capita cost to USS for workmen’s compensation. 138. Because the cost to USS is the best available source of the value to the plaintiffs of the fringe benefits, I find that the workmen’s compensation cost calculations in the Stipulated Damages Figures are a fair statement of the annual per capita value to plaintiffs of the workmen’s compensation payments. 139. USS did not offer any alternative method for computing workmen’s compensation benefits, which can be viewed either as a fringe benefit or a wage replacement (since Pennsylvania is a wage-loss state). In either case, this benefit is clearly part of this Court’s back pay calculation. 140. Using the tables from Stipulated Damages Figures with all USS employees as the reference group yields the following workmen’s compensation benefits for plaintiffs: a. Permanent Jobs Year of Hire Average Workmen’s Compensation Benefits 1972 ? 43,116 1973 238,224 1974 210,013 1978 271,765 1979 143,554 ALL YEARS TOTAL - $906,672 Summer Jobs Year of Hire Average Workmen's Compensation Benefits 1972 $ 608 1973 ■ 2,741 ALL YEARS TOTAL - $ 3,349 E. Unemployment Compensation Benefits 141. The Pennsylvania Unemployment Compensation law established an employer-financed program to replace wages lost by claimants who become unemployed through no fault of their own. 142. USS makes periodic unemployment compensation payments to the Commonwealth for its P & M employees at Fairless. Those payments go into a general fund established by the Commonwealth for payment of unemployment compensation claims. 143. The amount of compensation to which an employee is entitled is determined by the wages that he or she received while working. Because the hourly wage rate at USS is so high, its employees are entitled to the maximum rate of compensation. 144. USS argues that its contribution (cost) to the unemployment compensation fund should not be awarded to plaintiffs as part of their back pay recovery because these employer contributions are not direct payments to employees, and because the Third Circuit has held that any unemployment compensation benefits received by a Title VII plaintiff should not be deducted from a Title VII back pay award, as part of mitigation or “interim earnings” or “amounts earnable with reasonable diligence,” 42 U.S.C. § 2000e-5(g). Craig v. Y & Y Snacks, Inc., 721 F.2d 77, 83 (3d Cir.1983). 145. The Craig Court based this decision on the following characterization: Unemployment compensation most clearly resembles a collateral benefit which is ordinarily not deducted from a plaintiff’s recovery. 721 F.2d at 83. 146. However, simply because receipt of unemployment compensation is not deductible does not mean that unemployment compensation is not a benefit which plaintiffs are entitled to receive as a component of their total back pay award. 147. This is distinguishable from Craig, because here I am not concerned with deducting receipt of unemployment compensation benefits received, but rather, with predicting or estimating the increased value of unemployment compensation benefits as USS employees, over and above any unemployment compensation benefits plaintiffs received while employees of another company. 148. Thus, although unemployment compensation benefits most closely resemble collateral benefits, in order to make plaintiffs whole, plaintiffs are entitled to the difference between the unemployment compensation benefits they would have received at USS and those they did receive, if any, while employed outside USS. 149. The best way to measure this fringe benefit to plaintiffs is to use the cost to USS, with a percentage deduction for fringe benefits, as agreed upon by the parties. 150. Admittedly an imperfect measurement, it is the best available method of calculating the value of unemployment compensation benefits to the average plaintiff. 151. Further, use of the cost figures serves the dual purpose of making plaintiffs whole and penalizing the wrongdoer. Craig, supra, 721 F.2d at 83 and n. 2 (back pay “ ‘serves to work complete equity by penalizing the wrongdoer economically at the same time that it tends to make whole the one who was wronged’ ”) (emphasis supplied in Craig, quoting Franks v. Bowman Transportation Co., 424 U.S. 747, 787, 96 S.Ct. 1251, 1275, 47 L.Ed.2d 444 (Powell, J., concurring and dissenting)). 152. I find that plaintiffs are entitled to an award for unemployment compensation benefits, as measured by USS’s contribution to the unemployment compensation fund. 153. During any 52-week period of unemployment, an employee is entitled to an amount equal to 26 weeks of weekly unemployment compensation benefits, spread over the 52-week period. 154. A number of P & M employees hired by USS in the liability years were on layoff for at least some time since being hired. (Ex. P-209, No. 6). Had USS hired class members in the liability periods, they also would have experienced layoffs and would have received unemployment compensation benefits. 155. The amounts in the Stipulated Damages Figures are a fair representation of the per capita value to the plaintiffs of USS’s unemployment compensation payments. (Ex. P-200 at 78-90). 156. Using the tables from the Stipulated Damages Figures, with all USS employees as the reference group, yields unemployment compensation benefits for plaintiffs in the following amounts: a. Permanent Jobs Year of Hire Unemployment Compensation Benefits 1972 $ 42,032 1973 236,626 1974 201,473 1978 219,292 1979 108,305 ALL YEARS TOTAL —$807,728 b. Summer Jobs Year of Hire Unemployment Compensation Benefits 1972 $ 608 1973 9,595 ALL YEARS TOTAL -$ 10,203 F. Supplemental Unemployment Benefits (SUB) 157. SUB is a contractually required payment that USS makes to its employees on layoff to supplement what they receive from the Commonwealth in unemployment compensation. 158. USS employees hired in the liability periods are entitled to receive SUB from USS for up to 52 weeks, after a two-year waiting period. 159. Mr. Hensarling acknowledge that SUB represents a valuable fringe benefit to a USS employee. 160. Had plaintiffs been hired, some of them would have received SUB benefits. 161. The amounts in the Stipulated Damages Figures are a fair representation of the per capita value to the plaintiffs of USS’s SUB payments. (Ex. P-200 at 91-103). 162. USS contended, through the testimony of Mr. Hensarling, that the SUB amounts for the first two years for every cohort should be subtracted from the stipulated damages calculations to reflect a threshold requirement — namely that USS employees must have completed two years of continuous service before they are eligible for SUB. 163. I reject this adjustment. Mr. Hensarling admitted on cross-examination that his adjustment was overinclusive. His adjustment deleted not only the costs to USS for two years of new hires, but also costs for those USS employees eligible for SUB benefits. 164. Faced with a choice of under or over inclusiveness of damages, once again I find that defendant, as the wrongdoer, must suffer the loss, particularly because USS was in the best position to provide the most accurate data possible to reflect its cost of providing these fringe benefits to the plaintiff class. 165. Using the tables from the Stipulated Damages Figures, with all USS employees as the reference group, yields SUB benefits for the plaintiffs in the following amounts: Permanent Jobs Year of Hire SUB Benefits 1972 $ 45,134 1973 257,578 1974 214,296 1978 211,296 1979 96,642 ALL YEARS TOTAL — $824,946 Summer Jobs Year of Hire SUB Benefits 1972 $ 1,063 1973 9,595 ALL YEARS TOTAL — $ 10,658 VI. Mitigation of Wages and Fringe Benefits A. Mitigation Concept 166. The parties recognized that in determining damages under section 706(g) of Title VII, the amounts of lost USS wages and fringe benefits which 386 plaintiff class members could have earned at USS must be mitigated by amounts those persons subsequently earned, or with reasonable diligence could have earned, working at other jobs. 167. Defendant bears the burden of demonstrating the plaintiff class members’ interim earnings, or amounts eamable with reasonable diligence. 42 U.S.C. § 2000-5(g). 168. All uncertainties or doubts should be resolved in favor of the plaintiffs, because of Title VII’s dual goal: to make plaintiffs whole and to discourage employees from engaging in discriminatory acts. 169. Because it is impossible to ascertain which class members would have been hired by USS to fill the 386 jobs in question (of the approximately 13,000 class members), the efforts those 386 persons made to find employment elsewhere are unknown. Thus, the parties agreed to value the plaintiffs’ mitigation efforts by developing profiles of the class members and use published data gathered by the government to compute the earnings of similar blacks living in the eight county region known as the Philadelphia Standard Metropolitan Survey Area (SMSA). The outside wages obtained from the census data are then to be subtracted from USS lost wages to yield the plaintiffs’ net lost wages. 170. Both parties presented the testimony of highly qualified experts, who conducted mitigation studies for the class by developing profiles of the class members, projecting outside earnings, and then computing net damages. 171. The court finds that it is appropriate to address mitigation by estimating outside wages on a class-wide basis as follows: (a) by developing a profile of the class members in each liability period, based upon their age, sex, and educational traits because these traits, in combination, are considered statistically significant as earnings predictors; and (b) by using those age, sex, and educational profiles to obtain estimates of black earnings from local labor force census information; and (c) by basing such estimates upon earnings of blacks in the SMSA; and (d) by accounting in some reasonable manner for persons who made no attempt to find work comparable to that offered by USS; and (e) by accounting in some reasonable manner for inflation and real wage growth of the class members (including aging of the plaintiff class). 172. To mitigate fringe benefits, the parties stipulated to a formula for estimating the fringe benefits that the Black Hiring Shortfall would have received in the outside labor market. The ratio of outside wages to USS wages is applied to USS’s cost of fringe benefits, yielding outside fringe benefits. The outside fringe benefits amount is then subtracted from USS’s cost of fringe benefits, yielding mitigated fringe benefits. (Ex. P-200, Section V). B. The Killingsworth Studies 173. Plaintiffs’ expert, Dr. Mark Killingsworth, a labor economist, performed a series of analyses which satisfied the mitigation framework described above. He submitted an expert report detailing his studies (Ex. P-234), and he testified at trial. (1) Determination of USS Lost Wages 174. Using the Stipulated Damages Figures, Dr. Killingsworth selected the average annual wages of white employees who were hired by USS during the liability periods as the appropriate measure of the wages that the class members would have received from USS had they been hired. His report also contained similar calculations using the earnings of all USS employees. 175. As the Court previously found, use of “all” USS employees is the proper reference group for determination of what plaintiffs would have earned at USS, had they been hired. 176. For those plaintiffs who should have been hired in each liability period, Dr. Killingsworth made annual adjustments to the USS wages to account for attrition. (2) Developing Class Member Profiles 177. Dr. Killingsworth prepared class member profiles for each liability period. In doing so, he assumed that class members in any year possessed the characteristics and traits of those blacks that USS actually hired during that liability period. He also assumed that the traits that would be significant in determining outside earnings for the class would be age, sex and educational level. He did not assume that the traits of class members in one liability period would be the same as traits of class members in other liability periods. 178. In developing each of his five class member profiles, Dr. Killingsworth used information gathered by Dr. Litwin, an expert statistician who testified for plaintiffs at both the liability and damages portions of the trial. 179. Dr. Litwin made a random sample of employment applications submitted by blacks who were hired for P & M labor jobs at Fairless Works during each liability period. Dr. Litwin sampled the applications of approximately 530 black individuals who were hired by USS. 180. Dr. Litwin converted the following information obtained from the sample applications into machine-readable form: the sex/race code, the date of application, the date of hire, the year of birth, the number of years of education and the type of work requested. 181. From this computer-coded information, Dr. Litwin created cells, according to age, education, date of application, and year of hire. 182. Using Dr. Litwin’s data base, Dr. Killingsworth developed five separate profiles of the blacks who were actually hired by USS in the years 1972, 1973, 1974, 1978 and 1979. 183. Dr. Killingsworth decided to use the profiles of the USS hires (rather than ones based upon unsuccessful applicants) for three main re