Heterogeneity in long-term returns to eduction: An inconvenient truth

This paper studies the long-term relationship between parental and child education in Germany, where children are tracked into academic and non-academic track schools at the age of 10. On average, children are more likely to attend an academic track school if their parents attended one. Estimating marginal treatment effect curves, we find that there is no effect for disadvantaged individuals, suggesting that educational policies attempting to improve the educational prospect of disadvantaged individuals may fail to reduce inequalities in the long run. Low labor market returns despite better education is the main explanation for the null effect for these individuals.