Identification in Endogenous Sequential Treatment Regimes

This paper develops a novel nonparametric identification method for treatment effects in settings where individuals self-select into treatment sequences. I propose an identification strategy which relies on a dynamic version of standard Instrumental Variables (IV) assumptions and builds on a dynamic version of the Marginal Treatment Effects (MTE) as the fundamental building block for treatment effects. The main contribution of the paper is to relax assumptions on the support of the observed variables and on unobservable gains of treatment that are present in the dynamic treatment effects literature. Monte Carlo simulation studies illustrate the desirable finite-sample performance of a sieve estimator for MTEs and Average Treatment Effects (ATEs) on a close-to-application simulation study.