Bayesian Fuzzy Regression Discontinuity Analysis of the Effect of Education on Earnings (joint with Siddhartha Chib)

Given the endogeneity of the schooling decision a common approach is to exploit policy changes in the context of 2SLS approaches to estimate earnings effects of schooling. In this paper we use the 1947 increase in minimum school leaving age and the discontinuity in schooling intake it induced to find the causal effect of an extra year of schooling on earnings. We approach the problem from the perspective of a fuzzy regression discontinuity design, modelling the fuzziness in terms of a discrete latent type variable. We propose a Bayesian parametric framework to estimate the earnings effects for compliers