The Unequal Returns to Education: International Evidence from Bias-Corrected Quantile Regression
by Christopher J. Palmer
Ongoing globalization and technological change have affected labor markets around the world and the importance of education therein. This project will investigate the extent to which the returns to education vary across the income distribution and how different this pattern is across time and space. Recent research has used quantile regression methods to document a marked increase in the U.S. returns to education for the top quintile of the conditional wage distribution. We show that these estimated returns are downward biased because of errors in variables in the dependent variable—self-reported wages. We correct for this measurement error in a quasi-Maximum Likelihood approach and extend our analysis to a wide set of countries and years using World Bank household survey data to document that the income inequality contribution of the returns to education varies considerably across time and space.