Cross-Sectional Dependence Robust Block Bootstrap Panel Unit Root Tests

In this paper we consider the issue of unit root testing in cross-sectionally dependent panels. We consider panels that may be characterized by various forms of cross-sectionaldependence including (but not exclusive to) the popular common factor framework. Weconsider block bootstrap versions of the group-mean Im, Pesaran, and Shin (2003) and thepooled Levin, Lin, and Chu (2002) unit root coefficient DF-tests for panel data, originallyproposed for a setting of no cross-sectional dependence beyond a common time effect. Thetests, suited for testing for unit roots in the observed data, can be easily implemented asno specification or estimation of the dependence structure is required. Asymptotic propertiesof the tests are derived for T going to infinity and N finite. Asymptotic validity of thebootstrap tests is established in very general settings, including the presence of commonfactors and even cointegration across units. Properties under the alternative hypothesisare also considered. In a Monte Carlo simulation, the bootstrap tests are found to haverejection frequencies that are much closer to nominal size than the rejection frequenciesfor the corresponding asymptotic tests. The power properties of the bootstrap tests appearto be similar to those of the asymptotic tests.