This paper attempts to test whether information problems in labor markets help to explain why minority or female workers are sometimes paid less than equally-qualified white male workers. In particular, the relationship between starting wages, current performance, and race and sex is studied. OLS regressions of starting wages on current performance—which is measured some time after the beginning of employment—indicate that minority workers are paid lower starting wages than white workers with the same eventual performance, among both men and women. This could reflect taste discrimination.

However, if employers base starting wages on expected productivity or performance, and average performance is lower for minority workers (as it is in these data), then these estimated differentials could reflect simple statistical discrimination. Minority workers and white workers may each receive average starting wages equal to average performance, but a minority worker who turns out to be a high performer will end up getting a lower starting wage than a white worker who turns out to be a low performer, even if these workers turn out to have the same performance. A test of statistical versus taste discrimination, and a test of statistical discrimination versus pure measurement error, provide some evidence for both men and women that statistical discrimination is partly to blame for these differences in starting wages between minority and white workers, although the evidence is not very strong statistically.

Average performance of women in the sample studied in this paper is if anything higher than that of men, so simple statistical discrimination cannot explain the lower starting wages that women receive. However, more complex models of statistical discrimination suggest that worse labor market information about a particular group can lead to lower demand for that group (even conditional on the same average performance or productivity), and hence generate wage differentials.