A Test of the Worker Quality Explanation for the Urban/Non-Urban Wage Gap

We conduct a test to determine whether the observed urban/nonurban wage gap is a return to unobservable ability. A new measure of worker quality is added to wage regressions, and the estimated urban premium is compared with the results from the standard human capital measures. This new measure of unobserved ability refers to workers' literacy as assessed by the International Adult Literacy Survey (IALS). Using data from three countries with different population densities and scope of human capital spillovers, we find that controlling for literacy levels does not reduce the size of the urban wage premium, implying that the wage gap is not due to previously unobserved worker quality. We also test whether literacy affects interaction between the urban wage premium and experience, which may proxy for learning-by-doing, and find similarly insignificant changes in the estimates.