Heteroskedasticity robust panel unit root tests

This paper proposes new unit root tests for panels where the errors may be not only serial and/or cross- orrelated, but also unconditionally heteroskedastic. Despite their generality, the test statistics are shown to be very simple to implement, requiring only minimal corrections and still the limiting distributions under the null hypothesis are completely free from nuisance parameters. Monte Carlo evidence is also provided to suggest that the new tests perform well in small samples, also when compared to some of the existing tests.