Disparate impact realism.

• Author: Wax, Amy L.

TABLE OF CONTENTS INTRODUCTION AND SUMMARY I. DISPARATE IMPACT EMPLOYMENT DISCRIMINATION: THE DOCTRINE AND ITS UNCERTAINTIES II. EVIDENCE ON DISPARATE IMPACT: INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY (IOP) RESEARCH A. Job Screening Methods and Predictions of Job Performance B. Job Screening Methods and Adverse Impact: The Validity-Diversity Tradeoff III. SQUARING THE CIRCLE: COMPLYING WITH THE REQUIREMENTS OF DISPARATE IMPACT IV. REFORMING DISPARATE IMPACT: ALTER OR ABOLISH IT A. Alter It: Disparate Impact Realism B. Abolish It C. The Problem of False Negatives CONCLUSION INTRODUCTION AND SUMMARY

In Ricci v. DeStefano, (1) the Supreme Court recently reaffirmed the doctrine, first articulated by the Court in Griggs v. Duke Power Co., (2) that employers can be held liable under Title VII of the Civil Rights Act of 1964 (Title VII) for neutral policies with a disparate impact on minority workers. (3) The Court has further held that employers can escape liability by showing that a policy is job-related or consistent with business necessity. (4)

In the interim since Griggs, social scientists have generated a substantial body of research designed to help employers comply with the mandates of the doctrine. (5) This evidence has undermined two key elements of Griggs that have informed the application of the disparate impact rule more generally. First, Griggs and its progeny rest on the implicit assumption that fair and valid staffing practices will result in workers from each race being hired or promoted in rough proportion to their numbers in the background population or in an otherwise appropriately defined pool of candidates. (6) The so-called four-fifths rule, under which an employer is presumptively liable if the rate of hiring for minority workers is less than 80 percent of the rate for the majority white population, reflects this assumption. (7) Second, the Court in Griggs noted the absence of evidence that the screening criteria in that case--a high school diploma and scores on a "professionally prepared aptitude test[ ]"--were related to subsequent performance of the service jobs at issue, and expressed doubt about the existence of such a relationship. (8)

Social science research casts doubt on both of these aspects of Griggs. First, research in industrial and organizational psychology (IOP) has repeatedly documented that, despite their imperfections, tests and criteria such as those at issue in Griggs (which are heavily "g"-loaded and thus dependent on cognitive ability) remain the best predictors of performance for jobs at all levels of complexity. (9) Second, work in psychometrics, educational demography, and labor economics indicates that blacks, and to a lesser extent Hispanics, currently lag behind whites both in cognitive ability test performance and in the skills needed for success on the job. (10) These gaps are reflected in lower scores on the types of g-loaded job screens that best predict job success. (11) The combination of well-documented racial differences in cognitive ability and the consistent link between ability and job performance generates a pattern that experts term the "validity-diversity tradeoff": the most effective job selection criteria consistently generate the smallest number of minority hires. (12) Indeed, the evidence indicates that most valid screening devices will have a significant adverse impact on blacks and will also violate the four-fifths rule under the law of disparate impact. (13)

In sum, the IOP literature demonstrates that the empirical and demographic premises behind the disparate impact rule do not match reality and have turned out to be myths. (14) As a consequence, most legitimate job selection practices, including those that predict productivity better than alternatives, will routinely trigger liability under the current rule. (15) Although the Supreme Court in Griggs and subsequent cases has repeatedly stated that disparate impact doctrine is consistent with a rigorously competitive meritocracy, (16) employers seeking to maintain such a meritocracy among a diverse population will run a high risk of being sued for violations of the rule. (17) Such lawsuits will put employers to the onerous, uncertain, and sometimes impossible task of justifying their job selection practices. This may result in unwarranted liability or induce undesirable, self-protective strategies. (18) Even in the absence of those consequences, a proper application of the doctrine is unlikely to change the racial composition of the workplace or to increase demographic diversity. The best explanation for current workforce imbalances is the existence of real average group differences in knowledge, skills, and abilities. These human capital disparities, and not the use of non-merit-related selection or the erection of arbitrary barriers, best explain observed employment patterns. And given the present magnitude of skill differences and the shortage of qualified minority workers, the correct application of the disparate impact rule will not increase workforce diversity and could well make some jobs less diverse.

In light of these observations, this Article proposes to modify the doctrine of disparate impact to adopt a new standard of "disparate impact realism." (19) The disparate impact rule should be revised by making two changes in the standard that triggers potential liability. First, the target four-fifths ratio of minority to majority hires should be relaxed to reflect the empirically documented gap in actual productivity between whites and minority workers. Second, the fixed nature of the threshold ratio should be abandoned in favor of a sliding scale relationship, documented in the IOP literature, (20) that pegs expected group staffing patterns to measured disparities in group performance and the selectivity of particular positions. Although it does not altogether relieve employers of the burdens imposed by the disparate impact rule, disparate impact realism compares favorably with the current regime. By shrinking the number of employment practices that can potentially trigger liability, realism lessens the pressure to hire a racially balanced workforce, especially for highly selective jobs. Moreover, the uncertainties and potential constitutional difficulties generated by the Supreme Court's recent decision in Ricci v. DeStefano (21) make it desirable to cut down on the number of situations that can generate disparate impact claims. (22) Finally, disparate impact realism functions as an information-forcing device. (23) By making it easier for employers to satisfy the rule, and by aligning expectations with current labor demographics, realism enhances employers' incentives to devise personnel practices that maintain productivity while achieving maximally feasible diversity.

Alternatively, this paper proposes abolishing the disparate impact rule altogether. (24) The principal argument for repeal is that, under present social conditions, racial imbalances in employment are exceedingly weak evidence of discrimination, either in the form of race-based disparate treatment or through unlawful disparate impact. The IOP data indicate that the distribution of skill and human capital best explains observed patterns of adverse impact. The gaps by race in developed abilities, not race-based exclusion or arbitrary barriers to employment, are the principal factors behind racial imbalances on the job. (25) In light of these realities, the disparate impact rule is fatally overbroad and ensnares far too much conduct in its net. Under current social conditions, the vast majority of commonly used selection procedures are valid and job-related and thus do not actually violate the disparate impact rule. (26) Yet most valid personnel practices will routinely show enough adverse impact to create a prima facie case of discrimination, thereby shifting the burden of justification to employers. (27) Given the legal uncertainties and practical difficulties of defending disparate impact claims, employers run a significant risk of being found liable regardless of whether their defenses are valid and even though they are not actually violating the rule. (28) Because virtually no aspect of the business necessity defense is settled law, employers face the prospect of protracted, expensive, uncertain, and resource-intensive litigation to defend their practices. (29) This encourages them to engage in perverse, inefficient, and evasive tactics, including de facto affirmative action. (30) And even if employers avoid this response or successfully resist a disparate impact challenge, they may be required to expend significant time, attention, and resources on defending their personnel selection methods. In sum, the overbreadth of the disparate impact rule is both inefficient and fundamentally unfair. If properly applied, the rule will prove costly but do little or nothing to increase minority representation in the workforce. And the racial preferences that may result are directly at odds with the meritocratic goals of disparate impact principles.

The data currently reveal that most jobs are more diverse than disparate impact doctrine actually requires. (31) Indeed, blacks lag behind whites in performance on the job in many categories. (32) This indicates that employers are not arbitrarily excluding minorities from the workforce but are rather bending over backwards to include them. In addition, disparate impact litigation does nothing to correct the underlying skill deficits reflected in these on-the-job gaps and drains resources from that task. (33) The doctrine represents a costly, misplaced effort that could better be directed at addressing the root causes of workforce racial imbalance.

1. DISPARATE IMPACT EMPLOYMENT DISCRIMINATION: THE DOCTRINE AND ITS UNCERTAINTIES

   In Griggs v. Duke Power Co., (34) the Supreme Court ruled for the first time that job requirements with a disparate impact on minorities, despite...