Title: Testing for multiple structural breaks in multivariate long memory time series
Authors: Kai Wenger - Institute of Statistics (Germany) [presenting]
Philipp Sibbertsen - University of Hannover (Germany)
Simon Wingert - University of Hannover (Germany)
Abstract: Estimation and testing of multiple unknown breaks in multivariate long-memory time series is considered. We propose a likelihood ratio based approach for estimating breaks in the mean and the covariance of a system of long-memory time series. The limiting distribution of these estimates as well as consistency of the estimators is derived. A test to determine the unknown number of break points is given based on sequential testing on the regression residuals. A Monte Carlo exercise shows the finite sample performance of our method. We illustrate the usefulness of our test by analysing two real data sets.