The Returns to Education and the Determinants of Race and Gender Earnings Gaps Using Self-Reported versus Employer-Reported Earnings Data

Abstract

There are wide and persistent gender and race differences in wage and salary incomes. Much of the literature attempting to explain these differences relies on self-reported earnings. Many self-reported wages and salary incomes differ, however, from those reported by employers. These errors can bias the measurement of the returns to education and the decomposition of racial and gender earnings gaps. This article examines the implications of using selfreported data for measuring gender and race disparities in earnings. We find that relative to employer-reported data, self-reports understate the wage and salary incomes of black women with low educational levels and overstate the returns to higher education for black women. These previously unexplored and unrecognised findings of bias in the estimation of the returns to black women’s education and the large understatement of earnings among less-well educated black women have implications for interpreting the sources of race and gender earnings inequality.