Testing the null of stationarity against the alternative of a unit root: an application to the Italian post-war economy

Abstract

Testing stationarity of economic time series has become a central issue in empirical economics. However, the testing procedure commonly employed, i.e. the augmented Dickey - Fuller test for a unit root (ADF test), has become the subject of rising criticism. Recently, some authors have proposed to 'revert the burden proof' by testing the null hypothesis of stationarity. Following this new line of research, this paper compares, using a set of annual time series of the post-war Italian economy, the results of the ADF test and the test recently proposed by Kwiatkowski et al. (KPSS test), where the null hypothesis is one of stationarity. A procedure which consists of combining the answers of the two tests is also considered. The evidence shows that while the ADF test strongly supports the unit root hypothesis, either the KPSS or the combined procedure provide a much less definite conclusion.