Lasso for Instrumental Variable Selection: A Replication Study
复制了Belloni等人使用Lasso进行工具变量选择的模拟研究，并分析了惩罚参数选择的重要性，对处理大p小n问题的经济学者有参考价值。
Summary Recently, Lasso methods have been applied to economic questions. In a seminal paper, Belloni et al . ( Econometrica ; 80 (6): 2369–2429) make use of (post‐)Lasso for instrumental variable selection in a setting where the number of instruments p is large or might even exceed the number of observations n —a situation which is prevalent in many current applications. We replicate their simulation study with the statistical package R (R Development Core Team ( )) and, moreover, analyze in more detail the importance of the choice of the penalization parameter, a crucial component in applications. Copyright © 2015 John Wiley & Sons, Ltd.