A test of independence under local stationarity based on the local characteristic function




Faculty/Professorship: Mathematics for Business and Economics  
Author(s): Brunotte, Guy-Niklas
Year of publication: 2022
Pages: 1- 98
Language(s): English
Licence: Creative Commons - CC BY-NC-ND - Attribution - NonCommercial - NoDerivatives 4.0 International 
DOI: 10.13140/RG.2.2.36779.31523
Abstract: 
The present work introduces a consistent, asymptotic level alpha test that examines whether two locally stationary processes are asymptotically independent at the same point in time. Thereby, the test procedure is obtained by using a characteristic function-based test statistic and adapting the dependent wild bootstrap to the locally stationary situation. The performance of the introduced test is examined in two simulation studies and the test is applied to investigate whether stock yields, which are expected to originate from non-stationary processes, depend on each other.

The present work introduces a consistent, asymptotic level alpha test that examines whether two locally stationary processes are asymptotically independent at the same point in time. Thereby, the test procedure is obtained by using a characteristic function-based test statistic and adapting the dependent wild bootstrap to the locally stationary situation. The performance of the introduced test is examined in two simulation studies and the test is applied to investigate whether stock yields, which are expected to originate from non-stationary processes, depend on each other.
Keywords: Testing Independence, Local Stationarity
Type: Preprint
URI: https://fis.uni-bamberg.de/handle/uniba/55020
Release Date: 8. August 2022