The Performance of Structural Change Tests

Abstract

Abstract This paper provides analytical and Monte Carlo studies of the effect of different types of structural change (residual variance, process mean and process persistence) on the performance of the Chow/Wald stability test. We focus on the first-order autoregressive model, which has been used to estimate and assess changes in inflation persistence. Our results show that the autoregressive model is a difficult subject for the Chow test.