Meta-Learners for Estimation of Causal Effects: Finite Sample Cross-Fit Performance
This provides practical guidance for econometricians and researchers applying causal inference methods, though it is incremental as it focuses on optimizing existing meta-learners.
The paper studied the finite sample performance of meta-learners for estimating heterogeneous treatment effects, finding that sample-splitting and cross-fitting reduce bias and improve efficiency in large samples, while full-sample estimation works better in small samples.
Estimation of causal effects using machine learning methods has become an active research field in econometrics. In this paper, we study the finite sample performance of meta-learners for estimation of heterogeneous treatment effects under the usage of sample-splitting and cross-fitting to reduce the overfitting bias. In both synthetic and semi-synthetic simulations we find that the performance of the meta-learners in finite samples greatly depends on the estimation procedure. The results imply that sample-splitting and cross-fitting are beneficial in large samples for bias reduction and efficiency of the meta-learners, respectively, whereas full-sample estimation is preferable in small samples. Furthermore, we derive practical recommendations for application of specific meta-learners in empirical studies depending on particular data characteristics such as treatment shares and sample size.