DOUBLY ROBUST INSTRUMENTAL VARIABLE REGRESSION

Instrumental variable (IV) estimation typically requires the user to correctly specify the relationship between the regressors and the outcome to obtain a consistent estimate of the effects of the treatments. This paper proposes doubly robust IV regression estimators that only require the user to either correctly specify the relationship between the measured confounding variables (i.e., included exogenous variables) and the outcome, or the relationship between the measured confounding variables and the IVs. We derive the asymptotic properties of the doubly robust IV regression estimators and investigate their finite sample properties in a simulation study. We apply our method to a study of the effect of education on earnings.