Estimation and inference of treatment effects with L2-boosting in high-dimensional settings

被引:9
作者
Kueck, Jannis [1 ]
Luo, Ye [2 ]
Spindler, Martin [1 ]
Wang, Zigan [2 ,3 ]
机构
[1] Univ Hamburg, Fac Business Adm, D-20148 Hamburg, Germany
[2] Univ Hong Kong, HKU Business Sch, Hong Kong, Peoples R China
[3] Tsinghua Univ, Sch Econ & Management, Beijing, Peoples R China
关键词
L-2-boosting; Treatment effects; Instrumental variables; Post-selection inference; High-dimensional data; REGRESSION; SELECTION; REGULARIZATION; APPROXIMATION;
D O I
10.1016/j.jeconom.2022.02.005
中图分类号
F [经济];
学科分类号
02 ;
摘要
Empirical researchers are increasingly faced with rich data sets containing many controls or instrumental variables, making it essential to choose an appropriate approach to variable selection. In this paper, we provide results for valid inference after post- or orthogonal L-2-boosting is used for variable selection. We consider treatment effects after selecting among many control variables and instrumental variable models with potentially many instruments. To achieve this, we establish new results for the rate of convergence of iterated post-L-2-boosting and orthogonal L-2-boosting in a high-dimensional setting similar to Lasso, i.e., under approximate sparsity without assuming the beta-min condition. These results are extended to the 2SLS framework and valid inference is provided for treatment effect analysis. We give extensive simulation results for the proposed methods and compare them with Lasso. In an empirical application, we construct efficient IVs with our proposed methods to estimate the effect of pre-merger overlap of bank branch networks in the US on the post-merger stock returns of the acquirer bank. (c) 2022 Elsevier B.V. All rights reserved.
引用
收藏
页码:714 / 731
页数:18
相关论文
共 34 条
[1]   Sparse Models and Methods for Optimal Instruments With an Application to Eminent Domain [J].
Belloni, A. ;
Chen, D. ;
Chernozhukov, V. ;
Hansen, C. .
ECONOMETRICA, 2012, 80 (06) :2369-2429
[2]   Inference on Treatment Effects after Selection among High-Dimensional ControlsaEuro [J].
Belloni, Alexandre ;
Chernozhukov, Victor ;
Hansen, Christian .
REVIEW OF ECONOMIC STUDIES, 2014, 81 (02) :608-650
[3]   Least squares after model selection in high-dimensional sparse models [J].
Belloni, Alexandre ;
Chernozhukov, Victor .
BERNOULLI, 2013, 19 (02) :521-547
[4]  
Breiman L, 1998, ANN STAT, V26, P801
[5]  
Breiman L, 1996, MACH LEARN, V24, P123, DOI 10.1007/BF00058655
[6]   Boosting algorithms: Regularization, prediction and model fitting [J].
Buehlmann, Peter ;
Hothorn, Torsten .
STATISTICAL SCIENCE, 2007, 22 (04) :477-505
[7]  
Bühlmann P, 2006, J MACH LEARN RES, V7, P1001
[8]   Boosting with the L2 loss:: Regression and classification [J].
Bühlmann, P ;
Yu, B .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2003, 98 (462) :324-339
[9]   Boosting for high-dimensional linear models [J].
Buhlmann, Peter .
ANNALS OF STATISTICS, 2006, 34 (02) :559-583
[10]  
Chandrasekhar Arun., 2016, Econometrics of Network Formation