Presented By: Econometrics
Econometrics
Two-Step Estimation and Inference with Possibly Many Included Covariates presented by Xinwei Ma, University of Michigan and Bootstrap-Based Inference for the Maximum Score Estimator presented by Kenichi Nagasawa, University of Michigan
Two-Step Estimation and Inference with Possibly Many Included Covariates Abstract:
This paper develops resampling-based inference methods for two-step estimators with many included covariates, and applies these results to several treatment effect settings, including Marginal Treatment Effects, Local Average Response Functions, and IPW under unconfoundedness.
Bootstrap-Based Inference for the Maximum Score Estimator Abstract:
This paper develops new inference results for cube root consistent estimators employing the standard nonparametric bootstrap. The results give, in particular, the first valid inference method based on the standard nonparametric bootstrap for the Manski's maximum score estimator.
This paper develops resampling-based inference methods for two-step estimators with many included covariates, and applies these results to several treatment effect settings, including Marginal Treatment Effects, Local Average Response Functions, and IPW under unconfoundedness.
Bootstrap-Based Inference for the Maximum Score Estimator Abstract:
This paper develops new inference results for cube root consistent estimators employing the standard nonparametric bootstrap. The results give, in particular, the first valid inference method based on the standard nonparametric bootstrap for the Manski's maximum score estimator.
Related Links
Co-Sponsored By
Explore Similar Events
-
Loading Similar Events...