A1001
Title: Automatic debiased estimation with machine learning-generated regressors
Authors: Telmo Perez - University of the Basque Country (UPV/EHU) (Spain) [presenting]
Juan Carlos Escanciano - Universidad Carlos III de Madrid (Spain)
Abstract: Many parameters of interest in economics and other social sciences depend on generated regressors. Examples in economics include structural parameters in models with endogenous variables estimated by control functions and in models with sample selection, treatment effect estimation with propensity score matching, and marginal treatment effects. More recently, machine learning (ML) generated regressors are becoming ubiquitous for these and other applications such as imputation with missing regressors, dimension reduction, including autoencoders, learned proxies, confounders and treatments, and for feature engineering with unstructured data, among others. The first general method is provided for valid inference with regressors generated from ML. Inference with generated regressors is complicated by the very complex expression for influence functions and asymptotic variances. Additionally, ML-generated regressors may lead to large biases in downstream inferences. To address these problems, automatic locally robust/debiased GMM estimators are proposed in a general three-step setting with ML-generated regressors. The results are illustrated with treatment effects and counterfactual parameters in the partially linear and nonparametric models with ML-generated regressors. Sufficient conditions are provided for the asymptotic normality of the debiased GMM estimators, and their finite-sample performance is investigated through Monte Carlo simulations.
