Source-Country Income Inequality and Immigrant Self-Selection: A Difference-in-Difference Approach

Abstract

This paper revisits the income maximization hypothesis on immigrant self-selection. The traditional Roy model predicts that negative selection would arise when income distribution in the source country becomes more unequal relative to the destination country. However, the previous literature provides mixed evidence. Using data from the World Bank and U.S. Census between 2000 and 2010, we estimate a new specification that controls for country-specific fixed effects and unobserved global trends in immigration. The estimation results show that there exists no statistically significant relationship between immigrant skill composition and source-country income inequality, indicating that cross-sectional analysis suffers from omitted variable biases.