Some Properties of the Robust Trend Tests

Abstract

Formal testing of whether a time series contains a trend is greatly complicated by the fact that in practice it is not known whether the trend is embedded in an I(0) or I(1), series, that is, within a weakly or strongly autocorrelated series. In this article we would like to present the properties of behavior of the robust (to the order of integration of the data) trend tests of Bunzel and Vogelsang (2005), Harvey et al. (2007) and Perron and Yabu (2009). These statistics are termed ‘robust’ in the sense that the asymptotic critical values for testing hypotheses on the trend coefficient.