CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Tests for principal eigenvalues and eigenvectors Authors:  Jianqing Fan - Princeton University (United States)
Yingying Li - Hong Kong University of Science and Technology (Hong Kong)
Ningning Xia - Shanghai University of Finance and Economics (China)
Xinghua Zheng - HKUST (China) [presenting]
Abstract: CLTs are established for the principal eigenvalues and eigenvectors under a large factor model setting. As an application, we develop two-sample tests for difference in either the principal eigenvalues or principal eigenvectors. In particular, these tests can be used to detect structural breaks in large factor models. While there exist such tests, they can not distinguish between individual eigenvalues and/or eigenvectors. Our tests provide unique insights into the source of structural breaks.