Job assignment and promotion under statistical discrimination: evidence from the early careers of lawyers

Minorities continue to be severely underrepresented at the top levels of most occupations despite making dramatic gains in initial access to them. This fact is particularly striking in the legal profession where blacks are well represented in each associate class yet face significantly lower probabilities of making partner. To explain this divergence in the career paths of blacks and whites, I develop a dynamic model of statistical discrimination in which firms diversify their workforce by lowering the hiring standard for blacks. Despite such a diversity goal at hiring, task assignment and promotion decisions are not constrained by this policy. In this institutional setting, the model predicts that although blacks are more likely to be hired compared to observably similar whites, they are more likely to be placed in worse tasks and less likely to be promoted conditional on the same set of observables. However, conditional on task assignment, blacks and whites face similar promotion rates. I test the model's predictions using new data from the After the JD study -- a unique longitudinal survey tracking the professional lives of more than 4,000 lawyers. Compared to whites of similar credentials, blacks are much more likely to be hired into the best law firms. However, they are assigned to worse tasks and are less likely to be a partner. This black-white difference in promotion rates can be explained by quality differences in task assignments early in the associates' careers even controlling for measures of effort and career preferences. Results from this paper provide a unique explanation for the underrepresentation of minorities at the top of professional ladders by revealing how incompatible strategies in job assignment can reduce the number of minority promotions compared to the case without affirmative action.