B1926
Title: Uniform inference for local conditional quantile treatment effect curve under high-dimensional covariates
Authors: Xintao Xia - Iowa State University (United States) [presenting]
Abstract: Heterogeneous local quantile treatment effects are investigated for observational data with high-dimensional covariates, without relying on the strong ignorability assumption. Using a binary instrumental variable, parameters of interest are identified in a population subgroup (compliers) through a two-stage regression model. Lasso estimation is developed with a non-convex and non-smooth objective function to estimate these parameters and propose a de-sparsifying estimator for both pointwise and uniform inference. Moreover, uniform strong approximations to the local quantile treatment coefficient process are obtained by conditionally pivotal and Gaussian processes. Based on these strong approximations, bootstrap resampling methods are developed that can be used for constructing uniform confidence bands for the heterogeneous/conditional local quantile treatment effects given high-dimensional covariates. Finally, performance is evaluated through simulation studies.