Rank-based unit root testing in the presence of structural change under the null: simulation results and an application to US inflation

Abstract

The size distortion of the Dickey-Fuller (Journal of the American Statistical Association, 74, pp. 427–31, 1979) unit root test is examined in the presence of structural changes in both the level and variance of integrated time series. In contrast to previous studies, the empirically relevant situation in which such breaks occur simultaneously is examined. It is shown that the severe distortion observed for the Dickey-Fuller test can be dramatically reduced via application of a simple rank-based method. The simulation results presented are supported by an empirical examination of the integrated nature of US inflation where differing inferences are drawn using the Dickey-Fuller test and the rank-based Dickey-Fuller test.