Causal returns to education

Using 191 estimates from 145 studies across 54 countries, this paper synthesizes evidence on private returns to schooling. Quasi-experimental estimates imply an average causal return of about 10 % per year of education, with instrumental-variables (IV) estimates exceeding ordinary least squares (OLS) in roughly 80 % of cases. The IV-OLS gap is larger in lower- and middle-income countries. We interpret this pattern as reflecting high marginal returns in low-education environments and the fact that IV identifies local average treatment effects for individuals whose schooling is changed by policy variation. Schooling yields substantial earnings gains, especially when policies expand participation among constrained groups.