Abstract:We propose a new estimator for heterogeneous treatment effects in a partially linear model (PLM) with many exogenous covariates and a possibly endogenous treatment variable. The PLM has a parametric part that includes the treatment and the interactions between the treatment and exogenous characteristics, and a nonparametric part that contains those characteristics and many other covariates. The new estimator is a combination of a Robinson transformation to partial out the nonparametric part of the model, the Smooth Minimum Distance (SMD) approach to exploit all the information of the conditional mean independence restriction, and a Neyman-Orthogonalized first-order condition (FOC). With the SMD method, our estimator using only one valid binary instrument identifies both parameters. With the sparsity assumption, using regularized machine learning methods (i.e., the Lasso method) allows us to choose a relatively small number of polynomials of covariates. The Neyman-Orthogonalized FOC reduces the effect of the bias associated with the regularization method on estimates of the parameters of interest. Our new estimator allows for many covariates and is less biased, consistent, and $\sqrt{n}$-asymptotically normal under standard regularity conditions. Our simulations show that our estimator behaves well with different sets of instruments, but the GMM type estimators do not. We estimate the heterogeneous treatment effects of Medicaid on individual outcome variables from the Oregon Health Insurance Experiment. We find using our new method with only one valid instrument produces more significant and more reliable results for heterogeneous treatment effects of health insurance programs on economic outcomes than using GMM type estimators.

Instrumental variable estimation with observed and unobserved heterogeneity of the treatment and instrument effect: a latent class approach

Detecting Heterogeneous Treatment Effect with Instrumental Variables

Estimating Causal Effects of Discrete and Continuous Treatments with Binary Instruments

Uniformly Semiparametric Ecient Estimation of Treatment Eects with a Continuous Treatment

Testing for homogeneous treatment effects in linear and nonparametric instrumental variable models

Identifying Causal Effects of Discrete, Ordered and ContinuousTreatments using Multiple Instrumental Variables

Instrumental variable analysis in the presence of unmeasured confounding

Estimation of treatment effects under endogeneous heteroskedasticity

Local Effects of Continuous Instruments without Positivity

A semiparametric instrumental variable approach to optimal treatment regimes under endogeneity

Estimation of Heterogeneous Treatment Effects Using a Conditional Moment Based Approach

Exploring Heterogeneity in Clinical Trials with Latent Class Analysis

Estimating Heterogeneous Treatment Effects by Combining Weak Instruments and Observational Data

Instrumental variable estimation with first-stage heterogeneity

Nonparametric Bounds on Treatment Effects with Imperfect Instruments

Interpreting Instrumental Variable Estimands with Unobserved Treatment Heterogeneity: The Effects of College Education

Semiparametric Estimation of Treatment Effects in Observational Studies with Heterogeneous Partial Interference

Nonparametric identification and efficient estimation of causal effects with instrumental variables

Latent group structure in linear panel data models with endogenous regressors

Selective Machine Learning of the Average Treatment Effect with an Invalid Instrumental Variable

High-dimensional Model-assisted Inference for Local Average Treatment Effects with Instrumental Variables