Econometric Methods for Causal Inference

This course covers a modern approach to the estimation of causal treatment effects. Topics covered include the estimation of treatment effects under unconfoundedness, instrumental variables, regression discontinuity, and a range of panel data methods (diff-in-diff, synthetic controls, fixed effects). We also discuss nonparametric estimation techniques, including kernel regression and machine learning methods, focussing on how these can be incorporated into treatment effect estimators.