A1307
Title: Treatment effect estimation in high-dimension: An inference-based approach
Authors: Ulrich Aiounou - Aix-Marseille School of Economics (France) [presenting]
Abstract: Post-lasso and post-double lasso are becoming the most popular methods for estimating average treatment effects from linear regression models with many covariates. However, these methods can suffer from substantial omitted variable bias in finite samples. Autometrics, another variable selection method based on statistical inference, is considered and shown with simulation evidence that post-double autometrics performs well when the other methods fail and is illustrated in an application.
