核心概念
Nested NPIV estimation poses challenges due to ill posedness, but new estimators offer solutions with robust performance.
摘要
The content discusses the challenges of estimating nested nonparametric instrumental variable regression (NPIV) and introduces new estimators to address these challenges. The article outlines the theoretical framework, assumptions, and results of the proposed estimators. It also includes simulations demonstrating the robust performance of the new estimators across different data generating processes. The analysis focuses on achieving accurate estimation of causal parameters in short panel data models using proxy variables.
統計資料
"Several causal parameters in short panel data models are scalar summaries of a function called a nested nonparametric instrumental variable regression (nested NPIV)."
"Our nonasymptotic analysis has three salient features: (i) introducing techniques that limit how ill posedness compounds; (ii) accommodating neural networks, random forests, and reproducing kernel Hilbert spaces; and (iii) extending to causal functions, e.g. long term heterogeneous treatment effects."
"We measure long term heterogeneous treatment effects of Project STAR and mediated proximal treatment effects of the Job Corps."
引述
"Our main contribution is a theory of nested NPIV that prevents ill posedness from compounding in complex ways, and that is optimistic for causal inference."
"We provide what appears to be the first nonparametric estimation theory for nested NPIV in its full definition."
"Our proposals repeatedly outperform nested 2SLS in nonlinear, heterogeneous causal models using short panel data and proxy variables."