Multiway Cluster Robust Double/Debiased Machine Learning
Harold D. Chiang, Kengo Kato, Yukun Ma, Yuya Sasaki · Journal of Business and Economic Statistics · 2021
Harold D. Chianga*, Kengo Katob, Yukun Mac & Yuya Sasakica Department of Economics, University of Wisconsin-Madison, Madison, WIb Department of Statistics and Data Science, Cornell University, Ithaca, NYc Department of Economics, Vanderbilt University, Nashville, TNSupplementary materials for this article are available online. Please go to www.tandfonline.com/UBES.CONTACT Harold D. Chiang Chiang [email protected] Department of Economics, University of Wisconsin-Madison, William H. Sewell Social Science Building 1180 Observatory Drive Madison, WI 53706-1393.AbstractThis article investigates double/debiased machine learning (DML) under multiway clustered sampling environments. We propose a novel multiway cross-fitting algorithm and a multiway DML estimator based on this algorithm. We also develop a multiway cluster robust standard error formula. Simulations indicate that the proposed procedure has favorable finite sample performance. Applying the proposed method to market share data for demand analysis, we obtain larger two-way cluster robust standard errors for the price coefficient than nonrobust ones in the demand model.