State legislatures and the Equal Opportunity Employment Commission (EEOC) have moved in parallel in recent years to provide new protections for the employment prospects of some surprising groups: people who are unemployed, people who have poor credit, and people with past criminal convictions. These new protections confound our usual theories of what antidiscrimination law is about. These groups are disanalogous in a variety of respects to groups defined by such characteristics as race, sex, and national origin. But the legislators and regulators enacting these new protections were responding to pervasive problems they observed in the opportunity structure of our society--problems of a particular kind that I call bottlenecks. Essentially, these legal actors judged that poor credit, unemployment, and past criminal convictions were having too outsized an effect on a person's employment prospects. If many or most employers demand good credit, then good credit becomes a serious bottleneck: a narrow place through which workers must pass to reach a wide range of opportunities on the other side.
This Article argues that the anti-bottleneck principle--the principle that the law ought to ameliorate severe bottlenecks in the opportunity structure where it can feasibly do so--is not only a way of understanding these new, cutting-edge protections, but also a way of understanding much of the project of Title VII and our existing body of antidiscrimination law. This Article explores the role the anti-bottleneck principle plays in legislators ' decisions to enact antidiscrimination laws and in decisions by judges and by the EEOC about how to interpret and enforce such laws. The Article argues that the anti-bottleneck idea is at the heart of both disparate treatment law and disparate impact law--and that it should
cause us to think differently about the function of disparate impact law. The EEOC lawyers who started down the path that led to Griggs v. Duke Power understood that general ability tests were becoming a major bottleneck in the opportunity structure. By limiting the use of those tests, Griggs ameliorated a bottleneck that had arbitrarily constrained the opportunities of many whites as well as blacks.
Finally, turning from the positive to the normative, this Article defends the central--if previously unacknowledged--role that the anti-bottleneck principle plays in our law of equal employment opportunity. It is a profound challenge for any legal system to promote "equal opportunity " in a world of pervasive difference and inequality, where the mechanisms that perpetuate inequality shift over time. The anti-bottleneck principle turns out to be a strong and surprisingly practical response to these challenges.
Table of Contents Introduction I. The Challenge: Equal Opportunity and Employer Discretion II. Three New Kinds of Antidiscrimination Statutes--And Why They Are Here A. Credit Checks in Hiring B. "No Unemployed Need Apply" C. Ban the Box D. Ban the Box as Antidiscrimination Law III. The Anti-Bottleneck Principle A. The Anti-Bottleneck Principle and the Opportunity Structure B. The Relative Severity of Bottlenecks C. Situating Bottlenecks in the Opportunity Structure as a Whole: An Initial Example D. EEOC Enforcement Choices and the Anti-Bottleneck Principle IV. Griggs, Disparate Impact, and the Anti-Bottleneck Principle A. Bottlenecks and the Origins of Disparate Impact B. Griggs 'White Beneficiaries C. Opening Bottlenecks vs. Group-Based Redistribution D. Disparate Impact and Universal Remedies E. Disparate Impact and Equal Opportunity V. Bottlenecks and the Project of Antidiscrimination Law A. Why the Anti-Bottleneck Principle? B. Objections to the Anti-Bottleneck Principle C. Limits of the Anti-Bottleneck Principle D. Frontiers of the Anti-Bottleneck Principle Conclusion Introduction
In the past several years, American states and localities have enacted a new wave of employment discrimination statutes aimed at protecting the job prospects of people who are unemployed, who have poor credit, or who have past criminal convictions. (1) The Obama administration, members of Congress, and the EEOC have recently proposed parallel protections in each of these areas at the federal level. (2)
These new statutes are antidiscrimination laws. But they are antidiscrimination laws of a kind that most of our usual theories of antidiscrimination law are hard-pressed to explain. (3) People with criminal convictions, people who are unemployed, and people with poor credit are not groups for whom one would ordinarily expect our law to show particular solicitude. They are disanalogous in a variety of salient respects to groups defined by such characteristics as race, religion, sex, national origin, and age, the groups covered by Title VII and the ADEA. (Indeed it is not entirely clear that persons with poor credit constitute a "group" in any relevant pre-discrimination sense at all.)
Many of these new statutes also confound our usual ways of thinking about antidiscrimination law in another way. Most of the "ban the box" statutes about criminal convictions and most of the statutes about unemployment status do not actually bar discrimination on those grounds in an employer's final decision. Instead these statutes bar employers from erecting certain initial barriers that block the consideration of all such applicants--policies that "no unemployed need apply," or check-boxes on the initial application form asking applicants if they have ever been convicted of a crime (hence the name "ban the box"). An employer remains free to obtain the information at a later stage in the process and free to decide not to hire the applicant because of it. Such statutes define no "forbidden grounds" for employment decisions. Yet these statutes have an important practical effect. They ensure that an applicant can make it through an initial cut, giving her an opportunity to convince employers that perhaps, despite a past criminal conviction or a bout of unemployment, she is nonetheless the best candidate for the job.
Legislators enacted all these statutes in response to what they viewed as pervasive problems in the opportunity structure--problems of a particular kind that I call bottlenecks, (4) Essentially, legislators judged that poor credit, unemployment, or past criminal convictions were having too outsized an effect on a person's employment prospects because too large a proportion of employers either were using or might soon use these criteria to screen their applicants.
Imagine that in a labor market with numerous employers, just one decided to use credit checks to screen potential hires. In that case, there would be no significant bottleneck--and likely no calls for legislation. Plenty of job opportunities would remain open to those with poor credit history.
But now suppose credit checks plummet in price. Or suppose credit bureaus launch a marketing campaign and successfully persuade most employers to use their products to screen applicants. Now the good credit history requirement has become a serious bottleneck: a narrow place in the opportunity structure through which many people must pass if they hope to reach a wide range of opportunities that open out on the other side.
The more pervasive the use of a particular test, criterion, or practice across a wider range of job opportunities and firms, and the more strict or dispositive its effect on employment decisions, the more severe the bottleneck. (A bottleneck is even more pervasive if its effects extend beyond the employment sphere.) The new wave of antidiscrimination statutes with which I began have the purpose and the effect of ameliorating certain bottlenecks--that is, making them less severe--by making them either less pervasive or less strict or both.
So far, all this may seem an interesting but idiosyncratic tale of a few new cutting-edge statutes. But this Article argues that what I call the anti- bottleneck principle--the principle that the law ought to ameliorate severe bottlenecks in the opportunity structure where it can--is far more than that. It is a way of understanding a central dimension of the project of antidiscrimination law.
The anti-bottleneck principle is a way of understanding the function of the more familiar, paradigmatic antidiscrimination protections--laws against discrimination on grounds such as race and sex. Race and sex are among the most powerful bottlenecks in the opportunity structure of our society, in the sense that they have broad, pervasive effects, both direct and indirect, on everyone's opportunities. If it is not just a few employers, but lots of them, that evaluate applicants differently or steer applicants and employees into different roles based on race or sex--and especially if such effects are not confined to the world of employment, but extend as well to other domains such as education or housing--then it makes sense to use legal tools such as disparate treatment law and disparate impact law to make these bottlenecks less severe.
In the past several years, the EEOC has moved in parallel with state legislators to scrutinize employers' hiring decisions that turn on credit checks, unemployment, and past criminal convictions. (5) The EEOC's hearings, guidance, and enforcement actions regarding each of these employer practices focus on the practices' possible racial disparate impact. This emphasis reflects the agency's statutory charge: the EEOC's job is enforcing Title VII, not formulating new antidiscrimination legislation. The state legislators, as we shall see, approach the same problems from a different direction. They primarily emphasize not the racial disparate impact of these practices, but rather, their potential to create bottlenecks that many of their constituents--of all races--have difficulty passing through.
But in the end, these seemingly quite different approaches converge. Both the state legislatures and the EEOC invoke...