 |
Treatment effect estimation in high-dimension: An inference-based approach
1 : Aix-Marseille Sciences Economiques
(AMSE)
École des Hautes Études en Sciences Sociales, Aix Marseille Université, Ecole Centrale de Marseille, Centre National de la Recherche Scientifique, École des Hautes Études en Sciences Sociales : UMR7316, Aix Marseille Université : UMR7316, Ecole Centrale de Marseille : UMR7316, Centre National de la Recherche Scientifique : UMR7316
5-9 Boulevard BourdetCS 5049813205 Marseille Cedex 1 -
France
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 sample. We propose a new method called Post-Double-Autometrics, which is based on Autometrics, and show that this new method outperforms Post-Double-Lasso in some realistic situations.
