A standard object of empirical analysis in labor economics is a modified Mincer
wage function in which an individual’s log wage is specified to be a function of education,
experience, and an indicator variable identifying race. The hope is that estimates
can be informative about the impact of minority status on labor market success. Here
we examine the theoretical basis for this regression in a context in which individuals live
and work in different locations. Our theory calls into question the conventional approach,
which implicitly assumes that the “race wage gap” is a single parameter; instead, the
race wage gap is predicted to vary by location. With this insight in mind, we reevaluate
evidence about the black-white wage disparity in the United States.