Nonparametric Tests for Equality of Conditional Distributions（条件分布等式的非参数检验？）

主讲人：Xingyu Li(National School of Development, Peking University)

主持老师：（北大经院）王熙

参与老师：（北大经院）王一鸣、刘蕴霆、王法

（北大国发院）黄卓、张俊妮、孙振庭

（北大新结构）胡博

时间：2023年9月22日（周五） 10:00-11:30

地点（线下）：北京大学经济学院107会议室

报告摘要：

This paper proposes two easy-to-implement nonparametric tests for the equality of two conditional distributions. To avoid estimating conditional density functions, we transform the null hypothesis into an equivalent characterization. Based on two-sample U-process and weak convergence theory, we construct the Kolmogorov-Smirnov (KS) and Cramér-von Mises (CvM) statistics and find their asymptotic distributions. The critical values are constructed by a multiplier bootstrap method. The proposed KS and CvM tests are proved to be asymptotically size-controlled and consistent against each fixed alternative. A study on the local power of the tests is provided. Monte Carlo experiments illustrate good performance of the tests in finite samples.

主讲人简介：

Xingyu Li is currently a Ph.D. student at National School of Development, Peking University. His research fields include theoretical econometrics and applied econometrics.