Assimilation and the Earnings of Immigrants: New Evidence from Longitudinal Data

This paper presents fundamentally new evidence on the rate of economic assimilation by immigrants in the United States. Previous analyses based on the decennial Census are subject to bias due to nonrandom return migration. Using 40 years of longitudinal data on a nationally representative sample of natives and immigrants, I show that Census-based estimates of assimilation in earnings are biased upwards. This bias is due to less successful immigrants within an entry cohort leaving the U.S. over time, and the bias is substantial mainly for non-Hispanic white immigrants. Longitudinal analysis reveals that non-Hispanic white immigrants actually suffer earnings losses with time in the U.S., net of aging and time effects. Longitudinal data also strengthen the conclusion that more recent entry cohorts of non-Hispanic white immigrants have lower earnings than their counterparts who immigrated earlier.