African immigrants in the United States; returns to education; discrimination
This paper uses 2000 Census 5 percent Public Use Microdata Sample to investigate the relative earning pattern of immigrants from African countries, and explores the relevance of existing explanations of the low returns to education. The study uses the Extreme Bound Analysis to check the robustness of the variables of interests. The empirical findings from the conventional earnings regression conform to the theoretical expectations. However, not all the variables of interests are robust in Extreme Bound Analysis. This suggests that conventional specifications may not encompass all necessary information. Future study may explicitly controls for more detailed country-specific characteristics of the immigrant-sending countries.