Measurement Error, Misspecification, and the Return to Foreign Education

Abstract

I use unique data from the October Supplement of the Current Population Survey to show that the return to foreign education among immigrants is 3.3 percent. Previous studies generate upwardly biased estimates of this parameter because they (1) systematically misattribute domestic education as foreign education and (2) include domestic education as an endogenous control variable. The results indicate that foreign education is less portable than previously thought. Non-linear specifications indicate that the difference in the return to foreign education among immigrants and the return to domestic education among natives is limited to workers with less than twelve years of schooling.

JEL Classification: I21, J24, J31, J61