Subsampling change-point detection in persistence with heavy-tailed innovations

This paper considers how to detect structure change in persistence fromI(1) toI(0) with innovations in the domain of attraction of a κ-stable law. We derive the asymptotic distribution of test statistic and find that the asymptotic distribution of test statistics depends on the stable index κ which is often typically unknown and difficult to estimate. Therefore the subsampling method is proposed to detect changes without estimating κ. We establish the asymptotic validity of this method and assess its performance in finite samples by means of simulation study.