Opinion
No. 71 Civ. 2877 (RLC).
April 6, 2005
DEBEVOISE PLIMPTON, EDWIN G. SCHALLERT, PATRICIA G. CORLEY, JYOTIN HAMID, WENDY B. REILLY, New York, New York, Attorneys for Plaintiffs.
LAWYERS' COMMITTEE FOR CIVIL, RIGHTS UNDER LAW, MICHAEL L. FOREMAN, KRISTIN M. DADEY, AUDREY J. WIGGINS, Washington, DC, Attorneys for Plaintiffs.
PUERTO RICAN LEGAL DEFENSE AND EDUCATION FUND, ALAN LEVINE, JACKSON CHIN, New York, New York, Attorneys for Plaintiffs.
EDMUND P. D'ELIA, P.C., New York, New York, Attorneys for Defendant.
HIGHSAW, MAHONEY CLARKE, JOHN O'B. CLARKE, Washington, DC, Attorneys for Defendant.
OPINION
Plaintiff-Intervenors, for themselves and on behalf of a class of similarly situated individuals, together with the Equal Employment Opportunity Commission ("EEOC") ask the court to hold defendant Local 28 of the Sheet Metal Workers' International Association ("Local 28" or "the union") in contempt of court for violations of this court's Order and Judgment entered in 1975 ("OJ") and Amended Affirmative Action Program and Order ("AAAPO") entered by this court in 1983.
BACKGROUND
Familiarity with this more than thirty-year-old race discrimination suit is presumed. See EEOC v. Local 638, 401 F.Supp. 467 (S.D.N.Y. 1975) (Werker, J.), aff'd sub nom, EEOC v. Local 638, Local 28 of Sheet Metal Workers' Int'l Ass'n, 532 F.2d 821 (2d Cir. 1976); EEOC v. Local 638 Local 28 of Sheet Metal Workers' Int'l Ass'n., 1982 WL 445 (S.D.N.Y.) (Werker, J.), aff'd in part, rev'd in part, 753 F.2d 1172 (2d Cir. 1985), aff'd sub nom, Local 28 of Sheet Metal Workers' Int'l Ass'n v. EEOC, 478 U.S. 421 (1986); EEOC v. Local 638, Local 28 of Sheet Metal Workers' Int'l Ass'n, 889 F.Supp. 642 (S.D.N.Y. 1995) (Carter, J.), aff'd in part, rev'd in part sub nom, EEOC v. Local 638, 81 F.3d 1162 (2d Cir. 1996); EEOC v. Local 638-Local 28 of Sheet Metal Workers' Int'l Ass'n, 13 F.Supp. 2d 453 (S.D.N.Y. 1999) (Carter, J.), aff'd in part, rev'd in part sub nom, City of New York v. Local 28, Sheet Metal Workers' Int'l Ass'n, 170 F.3d 279 (2d Cir. 1999); EEOC v. Local 638 etc.; EEOC v. Local 28, 2001 WL 66327, at *2 — *3 (S.D.N.Y) (Carter, J.); Local 28 of the Sheet Metal Workers' Int'l Ass'n, 2003 WL 21767772 (S.D.N.Y.) (Carter, J.); EEOC v. Local 638 etc.; Local 28 of the Sheet Metal Workers' Int'l Ass'n, 2003 WL 21804837 (S.D.N.Y.) (Carter, J.); EEOC v. Local 638 etc.; Local 28 of the Sheet Metal Workers' Int'l Ass'n, 2004 WL 2414013 (S.D.N.Y.) (Carter, J.). Only those facts relevant to this decision are recited below.
This race discrimination suit was originally brought by government agencies against Local 28 on behalf of black and Hispanic members of the union. In 1975, this court found that Local 28 had discriminated against black and Hispanic union members in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. §§ 2000e et seq., and entered an order permanently enjoining Local 28 from discriminating in its recruitment or admission to the union and requiring the parties to work with a court-appointed administrator to establish an affirmative action program. Over the years, this court has issued a series of orders in an attempt to remedy Local 28's past discrimination and ensure its compliance with court-imposed remedies. Much of the litigation to date has resulted from Local 28's non-compliance with this court's orders after liability was established in 1975.
Local 28 has been held in civil contempt on three separate occasions for violating this court's orders. In 1982, Judge Werker, from whom the court inherited the case, held defendant in contempt for, among other things: adopting a policy of underutilizing its apprenticeship program to the detriment of nonwhites; refusing to conduct the court-ordered publicity campaign targeting nonwhite workers; issuing unauthorized work permits to white workers from sister locals to the detriment of nonwhite workers; and, failing to maintain and submit records and reports as required by court-order. See EEOC v. Local 638, 1982 WL 445 (S.D.N.Y.) (Werker, J.), aff'd in relevant part sub nom., 753 F.2d 1172 (2d Cir. 1985). The following year, in 1983, Judge Werker once again found Local 28 in contempt and adopted the AAAPO, which was affirmed by the Second Circuit, with modifications, and ultimately by the United States Supreme Court. See EEOC v. Local 638 Local 28 of Sheet Metal Workers' Int'l Ass'n, 1982 WL 445 (S.D.N.Y.) (Werker, J.), aff'd in part, rev'd in part, 753 F.2d 1172 (2d Cir. 1985), aff'd sub nom, Local 28 of Sheet Metal Workers' Int'l Ass'n v. EEOC, 478 U.S. 421 (1986). Then, in 1993, another round of contempt proceedings strikingly similar to the present one was initiated. The court ultimately found Local 28 in contempt — this time for the period from January 1, 1984 to March 31, 1991 — and entered in a series of remedial measures. See EEOC v. Local 638, Local 28 of Sheet Metal Workers' Int'l Ass'n, 889 F.Supp. 642 (S.D.N.Y. 1995) (Carter, J.), aff'd in part, rev'd in part sub nom., EEOC v. Local 638, 81 F.3d 1162 (2d Cir. 1996). Central to the court's 1995 contempt finding, and the subject of the current contempt motion, was the existence of a work hours disparity between white and nonwhite members of the union. Relying on the statistical analysis of court-appointed expert Dr. Bernard Siskin, the court found clear and convincing evidence of a continued statistically significant disparity between the hours worked by white and nonwhite members of Local 28. The court held that the disparity was attributable to Local 28's discrimination and that the union had "actively contributed to the inability of nonwhites to find work." Local 28, 889 F.Supp. at 663, aff'd in relevant part, 81 F.3d at 1174.
In 2001 and 2002, government plaintiffs and Local 28 entered into settlement negotiations that culminated in the submission of a proposed consent order. In response to the proposed consent order, the Hispanic Society of Local 28 and a group of individual black and Hispanic members of Local 28 filed a motion to intervene in this action as plaintiffs and objecting to the proposed consent order. By decisions dated July 30, 2003, and August 6, 2003, respectively, the court granted the motion to intervene and rejected the proposed consent order. See EEOC v. Local 638 etc.; Local 28 of the Sheet Metal Workers' Int'l Ass'n, 2003 WL 21767772 (S.D.N.Y.) (Carter, J.); EEOC v. Local 638, 2003 WL 21804837 (S.D.N.Y.) (Carter, J.). On October 27, 2004, the court granted plaintiff-intervenors' motion for class certification on behalf of all black and Hispanic persons who are or were at any time since 1984 members, either as journeypersons or apprentices, of Local 28 and who are or were "underemployed" as compared to their white counterparts. See EEOC v. Local 638 etc.; Local 28 of the Sheet Metal Workers' Int'l Ass'n, 2004 WL 2414013 (S.D.N.Y.) (Carter, J.). It is these now certified plaintiff-intervenors who, in conjunction with the EEOC, have initiated the present contempt motion.
This court continues to use the term "underemployed" as it has been defined for the purposes of this litigation. Specifically, it refers to those "black and Hispanic union members and former members [whose] work hours were two or more standard deviations below what would be expected" if there were no discrimination. See EEOC v. Local 638-Local 28 of Sheet Metal Workers' Int'l Ass'n, 13 F.Supp.2d 453, 465 (S.D.N.Y. 1999) (Carter, J.)
The New York State Division of Human Rights and the City of New York, both plaintiffs in this litigation, submitted letter-briefs endorsing the present motion for contempt but reserving judgment on the remedies requested by plaintiff-intervenors and the EEOC.
In the current motion before the court, plaintiff-intervenors and the EEOC (hereinafter "Plaintiffs") ask the court to hold Local 28 in contempt for continued violations of the OJ and the AAAPO for the period starting April 1, 1991, through May 2, 2003. Specifically, plaintiffs claim that Dr. Siskin's new regression analyses for the 1991 to 2003 period demonstrate that statistically significant work hours disparities between white and nonwhite workers persist, that the disparities are attributable to race and that Local 28 has failed to make diligent efforts to remedy these disparities.
The original motion for contempt was filed against both Local 28 and the Joint Labor-Management Apprenticeship Committee (the "JAC"), which operates the apprenticeship school. On November 22, 2004, the court approved a stipulation and order submitted by plaintiffs and the JAC withdrawing and dismissing the motion for contempt against the JAC. The JAC agreed to collaborate with plaintiffs and the special master to resolve all issues between them arising out of the contempt motion.
DISCUSSION
Procedural Arguments
As a preliminary matter, Local 28 raises a number of procedural defenses, including that plaintiffs' motion is one for criminal, and not civil, contempt and that it is time-barred under various theories. Each of these defenses is essentially a rehashing of procedural arguments that have previously been considered and rejected by this court and the Second Circuit in conjunction with the last contempt motion. The back pay remedy Local 28 claims is criminal in nature is precisely the same remedy we ordered, and the Second Circuit affirmed, during the 1995 contempt finding. See Local 28, 170 F.3d at 284 (affirming, as a proper remedy for civil contempt, district court's back pay remedy and procedures for determining individual back pay awards, and rejecting Local 28's argument that the remedy was actually a punitive remedy for criminal contempt). Moreover, the current contempt motion is neither barred by the statute of limitations nor the doctrine of laches. Plaintiffs are not commencing an action under Title VII but rather bringing a motion for contempt based on Local 28's disregard of court orders, which is, under Second Circuit law, an equitable proceeding "subject to equitable defenses and not legal defenses such as the statute of limitations." Brennan v. Nassau County, 352 F.3d 60, 63 (2d Cir. 2003) (refusing to apply Title VII statute of limitations to a motion to enforce a consent decree that had been entered under Title VII). Furthermore, plaintiffs did not unreasonably delay in raising their claims that Local 28 was in contempt but rather moved for intervention within one month of receiving notice that they had a legally protectable interest that was not being adequately protected. See Local 28, 2003 WL 21767772, at *1. The court thus rejects Local 28's procedural defenses and examines the allegations of contempt.
Liability for Contempt
A party may be held in contempt only if it is proven by "clear and convincing" evidence that the party violated a "clear and unambiguous" order of the court. Local 638, 81 F.3d at 1171 (quoting United States v. Local 1804-1, 44 F.3d 1091, 1096 (2d Cir. 1995)). The violation need not be willful, but it must be demonstrated that "the contemnor was not reasonably diligent in attempting to comply." Id. (quoting Local 1804-1, 44 F.3d at 1096).
OJ and AAAPO Provisions
Pursuant to the OJ entered by Judge Werker in 1975, Local 28 is permanently enjoined from taking any action which would "deprive or tend to deprive any individual of employment opportunities with Local 28 contractors. . . ." OJ ¶ 1. Specifically, the union is barred from "fail[ing] or refus[ing] to refer any individual for employment with sheet metal contractors," id. ¶ 1, and the OJ mandates that "Local 28 . . . shall provide nonwhite journeymen, and apprentices of Local 28 with the same assistance, including the assistance of Local 28's officers and business agents, in obtaining employment as that provided to white members and apprentices of Local 28." Id. ¶ 21(g). Furthermore, the AAAPO requires that "all members . . . of Local 28 share equitably in all available employment opportunities in the industry." AAAPO ¶ 3.
Local 28 does not dispute — nor could it — that these provisions of the OJ and the AAAPO are clear and unambiguous. See Local 28, 81 F.3d at 1171 ("the union does not seriously contend that either the OJ or the AAAPO was unclear or ambiguous, nor would such a claim based on the fact that they proscribed discrimination generally, have any merit.").
Work Hours Disparity Plaintiffs' Evidence
Plaintiffs contend that there is a statistically significant disparity between hours worked by similarly situated white and nonwhite journeypersons in violation of Local 28's obligation to provide equal work opportunities to its members as mandated by OJ ¶¶ 1 and 21(g) and AAAPO ¶ 3. As proof of defendant's continued contempt, plaintiffs point to Dr. Siskin's new set of statistical analyses for the years 1991 to 2003. Based on data provided by Local 28 and the special master, Dr. Siskin conducted a series of analyses from which he concluded that, for each year of the relevant period, nonwhite journeypersons worked statistically significant fewer hours than similarly situated white journeypersons. See Siskin Report, Table 2A. Collectively, Dr. Siskin found that nonwhite journeypersons worked 1,824,939 fewer hours during the 1991 through 2002 period than would have been expected absent discrimination. See id., Table 10A. His report indicates that in many years, the average shortfall per underemployed nonwhite journeypersons was in the hundreds (and in some years in the two-hundreds or even three-hundreds) of hours, and that for each year the collective shortfall for nonwhite workers was statistically significant at well over two units of standard deviation. See id., Table 2A. In 1991, the first year of the relevant period, the disparity represented 12.26 units of standard deviation, then dropped to a low of 3.71 units of standard deviation in 1999 before steadily increasing back up to 7.26 units of standard deviation in 2002. As the court noted in the past, "[a]s the number of standard deviations increases, the probability that a disparity occurred by chance decreases. When a disparity is represented by two or more units of standard deviation, there is less than a five percent chance that the variation occurred randomly, and social scientists and courts commonly deem the disparity statistically significant." Local 28, 889 F.Supp. at 660 (citations omitted); see also Hazelwood Sch. Dist. v. United States, 433 U.S. at 311 n. 17, 97 S.Ct. at 2743 n. 13 ("a fluctuation of more than two or three standard deviations would undercut the hypothesis that decisions were being made randomly with respect to race). In all, Dr. Siskin estimates the work hours disparity to represent approximately $110 million in potential wages lost due to discrimination during the time period. See Siskin Report, Table 10A.
2002 represents the last year for which a complete set of data was available. Based on data through May 1, 2003, the disparity in 2003 represented 8.24 units of standard deviation. Siskin Report, Table 2A.
In order to ascertain the role played by race and to ensure that the observed disparities were not the result of non-discriminatory factors, Dr. Siskin compared only "similarly situated" journeypersons. In defining what constitutes "similarly situated" union members, Dr. Siskin adopted the model suggested by the union's expert, Dr. Mann. Thus, Dr. Siskin ran regression analyses controlling for major variables which might affect hours worked, such as, seniority in Local 28, experience in sheet metal work, experience in other building trades, and the percentage of the year during which journeypersons were active in the union. This was substantially the same model used by Dr. Siskin in his prior calculation of damages following the 1995 contempt finding. Based on these regressions, Dr. Siskin concluded that non-discriminatory factors failed to explain the results of his analyses and that the work hour disparity was indeed race-based.
Regression analyses are statistical tools commonly used by social scientists and increasingly accepted by the courts to determine the effect of a number of factors ("the independent variables") on a particular phenomenon ("the dependent variable"). Ottaviani v. State Univ. of N.Y. at New Paltz, 875 F.2d 365, 366-7 (2d Cir. 1989), cert. denied, 493 U.S. 1021, 110 S.Ct. 721, 107 L.Ed.2d 740 (1990).
Dr. Siskin did not include two variables — residence and specialty in roofing or ventilation sheet metal work — in his current analyses, though they were included in those for the 1984-1991 period, because the union failed to provide him with data on them.
Moreover, Dr. Siskin addressed Local 28's contention that members who worked less than 420 hours a year, a predominantly nonwhite group, were not "actively" seeking work and should not be included in calculating the work hours disparity. Although Dr. Siskin disagreed with the union's premise that nonwhite journeypersons simply chose to work fewer hours than their white counterparts, he reran the hours regression and restricted it to those members who worked more than the union's minimum benchmark of 420 hours. Even assuming the union's premise to be correct, Dr. Siskin still found that statistically significant disparities between white and nonwhite journeypersons persisted for every year between 1991 and 2003, as well as over the entire time period.
Local 28's Defenses
Local 28 does not dispute that, as a whole, nonwhite union members worked less hours than their white counterparts. Instead, Local 28 contests the severity and the source of the disparity. Local 28's expert, Dr. Mann, seeks to discredit Dr. Siskin's results by both faulting his methodology and suggesting new non-discriminatory variables which might account for the observed work hours disparity. The court addresses both of these defenses in turn.
(a) Methodological Critiques
Dr. Mann's critique of Dr. Siskin's methodology is twofold. The first of these methodological critiques Dr. Mann terms the "confidence bound flaw." See Mann Critique, at 3-4. Dr. Mann faults Dr. Siskin for being overly inclusive in identifying individuals who should be entitled to make a claim for damages. Specifically, he would include only those individuals falling below the "prediction interval" while Dr. Siskin would include individuals falling below the "confidence interval." The court ruled on this point in favor of Dr. Siskin in the context of calculating damages following the 1995 contempt finding and sees no reason to reverse itself now. See Local 28, 2001 WL 66327, at *2 — *3. Furthermore, the confidence-bound flaw issue may be relevant when identifying specific individuals who might be potential claimants but not, as was Dr. Siskin's mandate here, when determining any overall differences between white and nonwhite union members as a group.
See EEOC v. Local 638, 2001 WL 66327, at *2 — *3 (S.D.N.Y) (Carter, J.) for a more detailed comparison of these approaches.
The second of Dr. Mann's methodological critiques is referred to as the "over inclusion flaw." See Mann Critique, at 4-5. Dr. Mann argues that his findings that both whites and nonwhites are roughly equally and likely to fall below the prediction interval — the extreme low end of the distribution of hours — demonstrate that nonwhites were not the victims of discrimination. Rather than diminish Dr. Siskin's conclusions, however, these findings actually bolster the claim that there continues to exist a widespread and consistent pattern of discrimination to the detriment of nonwhite union workers. Indeed, Dr. Mann's findings confirm that the work hour disparity is not driven by a handful of outliers — those at the very bottom end of the work hour curve — but rather results from an overall pattern which uniformly works to the disadvantage of nonwhites as a group. The court thus finds both these methodological critiques unavailing.
(b) Local 28's New Non-Discriminatory Factors
Dr. Siskin's methodology aside, Local 28's principal defense is that the disparity in work hours can be explained by a new set of non-discriminatory factors which relieve Local 28 of liability. See Sobel v. Yeshiva Univ., 839 F.2d 18, 34 (2d Cir. 1988) (party challenging the validity of a multiple regression analysis on the basis that it omitted factors must identify those factors and demonstrate that they would weaken the showing made by the analysis), cert. denied, 490 U.S. 1105, 109 S.Ct. 3154, 104 L.Ed.2d 1018 (1989). The thrust of Local 28's argument is that, while apprentices, nonwhite union members have poorer project scores and higher rates of tardiness and absenteeism than their white counterparts and that these marks reflect character traits contractors later take into consideration when hiring journeypersons. Local 28 bases its claim on data compiled by the Joint Labor Management Apprenticeship Committee and Trust ("JAC"), which operates the apprenticeship school. In 1988, the JAC began to maintain a database known as the Apprentice Training Information System ("ATIS"), which records information on apprentices while in the apprenticeship program, such as job history while an apprentice, including where they worked; their hours; whether they were discharged and, if so, why; their absenteeism and tardiness; and, their grades on projects and various tests. Using the ATIS data, Dr. Mann ran a new set of regression analyses based on the apprentice school records of nonwhites who were journeypersons in the period 1992 through 2002. Dr. Mann concluded that the new ATIS factors were explanatory of the work hour disparity, with test scores and attendance as the most important factors. Specifically, Dr. Mann believes that the ATIS factors account for up to 45% of the disparity for Hispanics and 20% of the disparity for blacks. Local 28 posits that it is appropriate to use the ATIS data to examine the disparity because the data measures factors (e.g., test scores, attendance) that are proxies for traits (e.g., productivity, work ethic) contractors validly take into account when exercising their hiring discretion. Thus, the union's argument goes, if nonwhites score poorly in these areas as apprentices, it is logical to assume they will continue to do so as journeypersons and that contractors will thereby hire disproportionately more white than nonwhite union members.
Before delving into the substance of Local 28's defense, the court notes that though Local 28 had a 30-year history of racial discrimination, began compiling the ATIS data more than 15 years ago, proposed the current variables used by Dr. Siskin close to 10 years ago and had never suggested the use of any other variables, including following Dr. Siskin's preliminary October 2002 report, the union nonetheless waited until less than two weeks before the November 22, 2004 hearing to proffer its explanatory regression analyses. Indeed, it did so despite Dr. Siskin's repeated assertions throughout that he would willingly incorporate into his analyses any additional data or variables Local 28 believed should be considered. That the original variables used by Dr. Siskin were first agreed to in the context of calculating damages following the 1995 contempt finding does not change the fact that Local 28 did not dispute their validity nor that it waited until a disturbingly late hour to introduce a whole new set of variables in an attempt to escape liability.
More importantly, the court remains unpersuaded by the union's attempts to explain away Dr. Siskin's results regarding the work hours disparity. Use of the ATIS variables is appropriate only if they validly predict performance and work as a journeyperson, a showing the court believes Local 28 has failed to make. Local 28 has never monitored the graduates of the apprentice program to analyze the school's success in forming journeypersons. Local 28's own fact-witness, the head instructor of the apprentice program, agreed that one could not "accurately predict, based on the student's performance, whether or not he or she will become a successful journeyperson." Tr. 388-89. Indeed, time spent at the apprentice program, about 4 weeks out of the year, represents only a small fraction of an apprentice's training, with the remainder coming through on the job training. Local 28 also fails to explain why ATIS factors like attendance and project scores have a negligible effect on work hours as an apprentice but a purportedly significant effect on future work hours as a journeyperson.
The ATIS variables themselves also yield nonsensical results, which undermine the data's predictive ability. For example, Dr. Mann's regressions for 1992 indicate that it was more advantageous, in terms of hours worked, to be absent 7 to 12 days than 1 to 3 days as an apprentice. Similarly, in 1994, journeypersons who scored between 81-90 on a test when they were apprentices worked fewer hours than journeypersons who scored below 80 on the same test when they were apprentices. The court is simply not convinced that these apprenticeship records are sufficiently predictive of future work as a journeyperson. Black journeyperson Robert Taylor's experience helps illustrate that success in the apprentice program does not predict success as a journeyperson. Although Mr. Taylor excelled in the apprentice program, he has testified to being chronically unemployed as a journeyperson for much of his time as a union member. Tr. 210-233.
Even assuming arguendo that test scores and attendance rates in the apprenticeship program could predict performance and work as a journeyperson — a position the court does not share — there are significant problems with the underlying data. As both experts agreed, a number of considerations — including whether the quality of instruction and mentoring received by white and nonwhite apprentices was equal, whether the apprentice school instructors marking apprentices absent or tardy did so in a fair or neutral manner and whether the tests themselves validly measure skills actually needed for work as a journeyperson — would have helped establish the legitimacy of using the ATIS data. Local 28 has failed to substantiate any of these points. Given the union's past history of discrimination, the court is understandably wary that seemingly objective data is in fact pretextual and a mask for discrimination. The "lay-off for cause" variable included in the ATIS data, for example, presumably reflects apprentice lay-offs that occurred prior to 1992 given that the data consists of the apprentice school records of nonwhites who were journeypersons starting in 1992. At least some of these lay-offs, therefore, presumably occurred in a time period during which Local 28 itself has argued that contractor discrimination was occurring and contributing to the work hours disparity.
Though, on the evidence presented, the court is satisfied that apprentice project scoring was conducted in a "blind" fashion.
In addition, because Dr. Mann's regression analyses looked only at journeypersons for whom complete ATIS data was available, the data covered only about 10% of the full relevant nonwhite population. This 10% subset contained in the ATIS file, however, generally had much smaller work hour disparities to start with than the rest of the union's nonwhite population. For example, in 2000 and 2002 while the baseline difference between Hispanic and white journeyperson hours in the non-ATIS sample was detrimental to Hispanics it was actually favorable to Hispanics in the ATIS sample. In effect, this meant that there was no disparity to explain away in Dr. Mann's analyses for those years even though the disparity continued to exist in Local 28's Hispanic population generally.
Dr. Mann's analyses of the ATIS data also contain a serious methodological flaw which further detracts from their reliability and the union's explanatory argument. In his regression analyses, Dr. Mann often includes race and/or its interaction as a variable in his model, even though doing so means that the model itself accounts for any differences between race. As a result, even where race is in fact a highly statistically significant factor, the effect that race has on the number of hours worked appears to be nil. For instance, Dr. Mann's own results show that in 1993 being black had an effect on hours worked of -344 hours, yet Dr. Mann's tables show a zero shortfall of hours for blacks in that year. The explanation for this seeming contradiction is that Dr. Mann's model adjusts a black journeypersons' results by 344 hours simply by virtue of their being black, thereby erasing any perceived shortfall. This is little more than a mathematical tautology whereby the results are skewed by incorporating discrimination directly into the model. Needless to say, these results do little to convince the court. Furthermore, the effect of some of the ATIS variables on the number of hours worked also differs according to whether a journeyperson is white or nonwhite. In fact, at least one of the ATIS variables predicts differently depending on race in every year of Dr. Mann's analysis. This, of course, runs contrary to the assumption that if a variable like tardiness matters, it should matter equally to whites and nonwhites. In light of these deficiencies, it is apparent that the ATIS data fail to detract from Dr. Siskin's results.
Finally, even if the ATIS variables were valid predictors and the data was reliable, regression analyses using those variables still demonstrate that nonwhite journeypersons suffer from a massive work hours disparity. Correcting for the serious methodological flaw in Dr. Mann's model, analyses run by Dr. Siskin using the ATIS variables showed that lost wages due to discrimination still totaled close to $80 million for nonwhite union workers for the 1991 to 2002 period. See Siskin Response Report, Table 1 and 2. The court thus remains unconvinced by Local 28's explanatory argument and finds that plaintiffs have established through clear and convincing evidence that nonwhite journeypersons continue to be deprived of employment opportunities as a result of their race in violation of this court's orders.
Local 28's Diligence
In its previous contempt decision, the court noted that the OJ and AAAPO impose, "an affirmative obligation upon the union to ensure that nonwhites are not deprived of equal employment opportunities. Consequently, even if the union is not solely responsible for the continuing failure of nonwhites to find work as journeypersons, it is liable for the disparity in hours unless it can show that it has made diligent efforts to ensure equal work opportunities for all its members." Local 28, 889 F.Supp. at 663, aff'd in relevant part, Local 28, 81 F.3d at 1175. Thus, the OJ and the AAAPO require Local 28 not only to refrain from actively discriminating against its nonwhite members, but also to take affirmative steps to ensure equal work opportunities for all its members. The court finds that Local 28 has failed to meet this obligation.
While the court welcomes the few indications that the union is moving towards compliance with its orders, they are patently insufficient in the face of Local 28's overall failure to attempt to remedy the work hours disparity. Local 28 offers a series of steps it alleges it has taken in furtherance of compliance. Many of the steps Local 28 places its greatest reliance on — including the furlough program, the target agreements and its meetings with contractors — were not, even according to Local 28's witnesses, intended to address the work hours disparity, but rather to address unemployment generally or to increase available hours for all members. Such steps do not reflect diligent efforts to ameliorate the disparity in the distribution of hours between whites and nonwhites. Local 28 has made no effort to monitor what effect, if any, such steps had on the work hours disparity. Absent evidence to the contrary, one might assume that any work opportunities that arose as a result of these steps were distributed disproportionately to white journeypersons. Some steps — such as providing two sensitivity training sessions in the 13-year contempt period and engaging in various charitable activities — are trivial on their face. Other steps — such as appointment of minorities, the voluntary referral hall and the decision to suggest a hiring hall after first successfully petitioning for its defeat in the Second Circuit — came only in the latter portion of the contempt period. Cf. Local 28, 889 F.Supp. at 667-68 (noting that "`a spur of activity at the heels of plaintiffs' motion for a finding of contempt' is not evidence of a diligent effort to comply") (quoting Aspira of New York, Inc. v. Bd. Of Educ. Of the City of New York, 423 F.Supp 647, 654 (S.D.N.Y. 1976) (Frankel, J.). Though the election of two nonwhite candidates to officer positions in 2004 is encouraging, the court cannot ignore the fact they are the first nonwhites ever to be elected to such positions in the union's history and the first to be endorsed by an incumbent president of the union. Taken together, Local 28's purported affirmative steps simply do not constitute sufficient diligent efforts to ensure compliance. The reality remains that the work hours disparity between white and nonwhite journeypersons has continued unabated and that the union has failed to show it has taken "all the steps in its power to try to reduce the hours disparity." Local 28, 81 F.3d at 1175.
Throughout this litigation, Local 28 has repeatedly been admonished for its discriminatory practices and been ordered to redress the situation of its nonwhite members. Dr. Siskin's current analyses now confirm that, at least through 2002, the union's nonwhite members continued to work disproportionally fewer hours than their similarly situated white counterparts. Local 28 did nothing to monitor whether the work hours disparity continued through the current contempt period though it was in possession of all the underlying data. No Local 28 official admitted to tracking the work hours disparity nor was union leadership aware whether any Local 28 official had ever been asked to do so. Instead, Local 28 seemed content to wait for the next contempt proceeding to find out how its nonwhite journeypersons were fairing. Unfortunately for the union, these post facto checkups arrive too late.
See John Harrington's, current Local 28's president (1998-present), testimony (Tr. 348).
In its 1998 opinion, the court made provisions for the appointment of a statistical expert to monitor the union's compliance with the OJ and AAAPO. Local 28, 13 F.Supp. 2d at 468-69. Dr. Siskin was eventually chosen for that position and pursuant to a request by special master David Raff compiled the aforementioned regression analyses.
Similarly, Local 28 has failed to adequately address the problem of business agents referrals. Central to the 1995 contempt order was the finding that Local 28's business agents played a critical role in helping journeypersons obtain work and that they used this power to discriminate against nonwhite members. See Local 28, 889 F.Supp. at 663. Until July 1998, however, Local 28 continued not to record business agents referrals nor did Local 28 put into place any mechanism to track when and whether Local 28 officers were making work referrals. Moreover, Local 28 failed to develop any guidelines, protocols, or other written materials setting forth any standardized procedures for making job referrals or keeping track of requests for referrals. Throughout the contempt period, Local 28 business agents enjoyed unreasonable discretion in referring work. This, despite the fact that prior court opinions condemned the discriminatory manner in which that discretion had been wielded in the past. The experience of Robert Taylor, a black journeyperson, helps illustrate the failures of the referral system. Throughout the contempt period, Mr. Taylor diligently sought referrals by calling and writing letters to business agents as well as union officers. Tr. 210-233. Despite these efforts, Mr. Taylor recalled obtaining only one job through a business agent referral since 1991. Tr. 250-251.
In 1998, the court ordered a "blanket ban" on business agents referrals because business agents continued to refer work in a discriminatory manner, and ordered the establishment of a voluntary referral hall. See Local 28, 13 F.Supp. 2d at 468. The voluntary referral hall took five years to establish during which time the ban on business agents referrals was stayed. As a result, business agents continued to play a critical role throughout the current contempt period.
Journeypersons looking for work are sometimes told to put their names on a "manpower list" but Local 28 does not require business agents or other officers to use the list to make referrals and has no policies or procedures outlining how referrals from the list should be made.
See Edward Wilton deposition at 72-75 (stating that he receives three to four calls everyday for work referrals, that he often gives priority to those he knows and that he does not keep any records of the calls he personally receives from journeypersons looking for work).
Local 28 makes much of the fact that approximately 60% of its work referrals since July 1998 — when the union started recording such matters — were made to nonwhite journeypersons. While the court welcomes any indication that the union has worked to better access to work opportunities for nonwhite union members, this statistic is far less meaningful than Local 28 holds it out to be. First, it is based only on post June 1998 data and thus fails to account for referrals during the first seven years — approximately three-fifths — of the current contempt period. Second, the referral data contains no information on work referrals requested by Local 28 journeypersons. As a result, the 40% of referrals going to whites may constitute 100% of the referrals requested by whites, while the 60% of referrals going to nonwhites may constitute only a small fraction of the requests for referrals by nonwhites. Indeed, fewer whites may need to request referrals because more whites are employed in the first place. Finally, Local 28's referral records are fraught with errors. For example, Local 28's proposed exhibit J, which purported to compile Local 28's records of business agents work referral data, misleadingly suggested that all business agent job referrals went to unemployed journeypersons. However, the underlying job referral reports reveal that business agents often referred jobs to journeypersons who were already employed, or that they failed to note the employment status of the journeyperson who received the job referral. Local 28's exhibit JJ, a compilation of the underlying business agents referral reports is also incomplete in material ways. A number of the reports are not filled out completely and Local 28's weekly spreadsheets compiling the business agents referral data often conflict with the information provided on the underlying forms. Local 28's list of persons who made referrals during the years 1991-2004 is also unreliable. For example, though the former president of the union testified to not making any referrals from 1993 to 1998, Local 28's list indicates that he did so in every one of those years. This is by no means the first time the court has taken issue with Local 28's record keeping. See Local 28, 889 F.Supp. at 652 ("Local 28 has a tradition of failing to provide and maintain accurate records as required by the OJ and the AAAPO") (citations omitted); see also Local 28, 81 F.3d at 1172 ("[w]e agree with the district court's characterization of Local 28's efforts at compliance during [the last contempt] period as `irresponsible and sloppy.' The errors here were not mere clerical mistakes, but instead tend to demonstrate that Local 28 has systematically attempted to undercount white membership and overcount nonwhite membership.") (citations omitted). In short, the court believes Local 28 failed to sufficiently address its business agent referral system in an effort to remedy the work hours disparity during the current contempt period.
See, e.g., Ex. JJ (revealing 70 examples of referral reports that fail to provide the estimated length of the job); see also Ex. 21 (compiling 44 examples of referral reports by business agent Wilton in which he failed to specify whether the journeyperson was unemployed at the time of the referral).
See, e.g., Ex. JJ (in which of 41 of the weekly spreadsheets for the year 2000, 18 contained information that did not match the information in the underlying referral forms).
Compare Tr. 293 (Casey Testimony), with Ex 20 (Local 28's list of persons who have referred members to jobs, revised by Local 28 on November 22, 2004).
In addition, Local 28 has done little to encourage its foremen and superintendents to carry out their functions in a non-discriminatory manner. Union members serving as foremen and superintendents are responsible for the supervision of workers and are often given hiring and firing authority by the contractors. Yet, Local 28 does not monitor which of its members are serving as foremen and superintendents nor which journeypersons those members have hired. Moreover, Local 28 has established no written guidelines, procedures or policies to encourage Local 28 foremen and superintendents to exercise that discretion in a non-discriminatory fashion.
Instead, Local 28 has shown continued passivity in the face of a continuing work hours disparity and an unwillingness to properly confront its court-ordered responsibilities. Local 28 has done little to address the operative provisions of the OJ and AAAPO and has failed to convince the court that it is committed to providing equal work opportunities. In his deposition testimony, Joseph Casey stated that, to his knowledge, during his tenure as president of Local 28 (1993-1998), Local 28 took no steps to ensure the union was complying with OJ ¶¶ 1 or 21(g), aside from possibly investigating complaints of discrimination. Casey Dep. at 51, 64. President Casey subsequently backtracked from this testimony at the contempt hearing. Though the union presented evidence of two instances where it fielded and acted on complaints of racial discrimination, testimony also established that concerns raised by nonwhite members were ignored. The current president of the union, John Harrington (1998 — present), for instance, testified in response to questioning from the court that although members had informed him that nonwhite journeypersons were not getting jobs on par with white journeypersons he had done nothing to investigate these concerns. Tr. 357-359. Similarly, business agent Edward Wilton stated in his deposition that though he received specific complaints from minority members who informed him that they were laid off and that white members were then hired for the same jobs he took no action in response because employers had the right to hire and fire as they saw fit. Wilton dep. at 151-152. Simply put, Local 28 must be more actively involved in combating the work hours disparity if it hopes to escape a finding of contempt.
Local 28's lack of engagement in rectifying discriminatory practices has had a severe detrimental effect not only on its nonwhite union members but also on the union as a whole. Local 28 has endured years of litigation, with all the attendant costs both to its financial resources and its reputation. For this, the union has only itself to blame. Perpetual resistance and foot-dragging on the union's part will not cow the court from using the full range of powers at its disposal to ensure that Local 28 fulfills its court-ordered responsibilities. Sadly, the union continues to evade its obligations and the court has no choice but to hold Local 28 in civil contempt, once again.
CONTEMPT REMEDIES
The court is mindful of its obligation to take into consideration the union's financial state in fashioning relief for contempt. See Local 28, 81 F.3d 1162, 1177 (2d Cir. 1996). The court hereby orders the union to submit to a comprehensive financial audit to be conducted by an auditor appointed by special master David Raff. The auditor will work under the supervision of the special master and in cooperation with the union to determine the overall financial health of the union as well as its ability to cover costs associated with court-ordered contempt remedies. The auditor's review shall be completed within 60 days of the issuance of this order. Upon completion of the review, the auditor shall submit copies of its report to the parties, the court and the special master. The parties and the special master will then have 30 days from receipt of the auditor's report to submit their recommendations regarding the remedies requested by plaintiffs. Taking into account these submissions, the court will then address what remedies, both coercive and compensatory, are appropriate given the present finding of contempt.
IT IS SO ORDERED