| A large fraction of the literature on program evaluation focuses on efficient, flexible estimation of treatment effects under the assumption of unconfoundedness. The first two chapters of this dissertation contribute to this literature by studying the efficient estimation of a large class of multi-valued treatment effects as implicitly defined by a collection of possibly over-identified non-smooth moment conditions when treatment assignment is assumed to be ignorable. Chapter 2 proposes two general estimators, one based on an inverse probability weighting scheme and the other based on the efficient influence function of the model, and provides a set of sufficient conditions that ensure root- n consistency, asymptotic normality and efficiency of these estimators. Chapter 3 shows that, under mild assumptions, these conditions are satisfied for the marginal mean treatment effect and marginal quantile treatment effect, two estimands of particular importance for empirical applications. Previous results for average and quantile treatments effects may be seen as particular cases of the methods proposed in Chapter 2 when treatment is assumed to be dichotomous. Chapter 3 also illustrates the empirical applicability of the results derived in Chapter 2 by studying the effect of maternal smoking intensity during pregnancy on birth weight. The main empirical findings suggest the presence of approximately homogeneous, non-linear treatment effects concentrated on the first 10 cigarettes-per-day smoked during pregnancy.;Finally, Chapter 4 derives the optimal rates of convergence for the Block Regression Estimator, a nonparametric estimator of the regression function that is implicitly used when estimating the Average Treatment Effect by subclassification on the propensity score. This result contributes to both the literature of program evaluation and the literature of nonparametric estimation. |
