The Effects of Additive Outliers on Tests for Unit Roots and Cointegration.

This paper discusses the properties of the univariate Dickey-Fuller test and the Johansen test for the cointegrating rank when there exist additive outlying observations in the time series. The authors provide analytical as well as numerical evidence that additive outliners may produce spurious stationarity. The results easily generalize to models with 'temporary change' outliers. Through an empirical example, the authors discuss how additive and temporary change outliers can be detected in practice and they show how dummy variables can be used to remove the influence of such extreme observations.