Maria Nareklishvili: Heterogeneity of Partially Identified Treatment Effects

This paper identifies and estimates heterogeneous treatment effect bounds based on bounded support of the outcome and the monotonic treatment selection assumption. Specifically, I propose multivariate random forests that can be used to fit the bounds of treatment effects identified as the solution to a set of local moment equations. To detect heterogeneous subgroups, multivariate random forests adaptively search for subsets of data that exhibit the highest variation in the treatment effect bounds. The large sample theory in this paper shows that the method recovers consistent and asymptotically normally distributed parameters. Simulation experiments and applications to the National Longitudinal Survey of Youth suggest significant heterogeneity in the effect of the Head Start program on years of schooling.