A0966
Title: Testing for multiple structural breaks in multivariate long-memory regression models
Authors: Paulo Rodrigues - Universidade Nova de Lisboa (Portugal) [presenting]
Abstract: Estimation and testing of multiple breaks is considered that occur at unknown dates in multivariate long-memory time series regression and allows for the possibility of a cointegrated system. A likelihood ratio-based approach is proposed for estimating breaks in the regression parameter and the covariance of a system of long-memory time series regression. The limiting distribution of these estimates as well as the consistency of the estimators are derived. A testing procedure to determine the unknown number of breakpoints is given based on iterative testing on the regression residuals. A Monte Carlo exercise shows the finite sample performance of the method. An empirical application to inflation series illustrates the usefulness of the procedures.
