This letter reports an empirical analysis of individuals' education as a function of their parents' education and of education in the census tract of their upbringing. The analysis is conducted with data from the Panel Study of Income Dynamics, including confidential geocoded data that links individuals to contextual data from the 1980 and 1990 US Census. The results validate those of Kremer's on the role of neighbourhood effects, but also confirm, using both parametric and nonparametric methods, that parents' education and neighbours' education have nonlinear effects that are broadly consistent with theoretical predictions.
Read full abstract