This paper has two themes, First, we classify some effects which outli
ers in the data have on unit root inference, We show that, both in a c
lassical and a Bayesian framework, the presence of additive outliers m
oves 'standard' inference towards stationarity. Second, we base infere
nce on an independent Student-t instead of a Gaussian likelihood. This
yields results that are less sensitive to the presence of outliers. A
pplication to several time series with outliers reveals a negative cor
relation between the unit root and degrees of freedom parameter of the
Student-t distribution, Therefore, imposing normality may incorrectly
provide evidence against the unit root.