以因果隨機森林估計分量處置效果

Abstract

根據 Athey、Tibshirani、與 Wager (2019, The Annals of Statistics) 所建構的一般化隨機森林架構，本文探討如何以因果機器學習的方法估計工具變數分量迴歸。我們提出的計量方法能無母數地估計分量處置效果，並且衡量各個控制變數在異質性上的重要性。本文也依據此計量方法重新檢視兩個實證研究: 401(k) 退休金制度對財富的處置效果、以及職業訓練對所得的影響。

We propose an econometric procedure based mainly on the generalized random forests of Athey, Tibshirani and Wager (2019, The Annals of Statistics). Not only estimates the conditional quantile treatment effect nonparametrically, but our procedure yields a measure of variable importance in terms of heterogeneity among control variables. We also apply the proposed procedure to reinvestigate the distributional effect of 401(k) participation on net financial assets, and the quantile effect of participating a job training program on earnings.