Wednesday Seminar: Semiparametric Instrumental Variables Method

About this event

This paper proposes to construct instrumental variables by projection and expansion. As the projection and expansion process is semiparametric, we define this procedure as a semiparametric instrumental--variable (SIV) method to address endogeneity in regression models without relying on externally supplied instruments. The approach constructs instruments directly from observed regressors through a projection-based decomposition of the structural error. The method applies to linear, nonlinear, and non-- and semi--parametric models and provides a simple and computable alternative to conventional instrumental variable approaches. This paper establishes identification conditions and derives the asymptotic properties of the resulting estimators. It then proposes a simple LASSO selection method to examine the finite-sample performance of both the proposed method and the established theory by simulated and real data examples.