Unit root tests for panel data

This paper develops unit root tests for panel data. These tests are devised under more general assumptions than the tests previously proposed. First, the number of groups in the panel data is assumed to be either finite or in finite. Second, each group is assumed to have different types of nonstochas tic and stochastic components. Third, the time series spans for the soups a re assumed to be all different. Fourth, the alternative where some groups h ave a unit root and others do not can be dealt with by the tests. The tests can also be used for the null of stationarity and for cointegration, once relevant changes are made in the model, hypotheses, assumptions and underly ing tests. The main idea for our unit root tests is to combine p-values fro m a unit root test applied to each group in the panel data. Combining p-val ues to formulate tests is a common practice in meta-analysis. This paper al so reports the finite sample performance of our combination unit root tests and Im et al.'s [Mimeo (1995)] t-bar test. The results show that most of t he combination tests are more powerful than the t-bar test in finite sample s. Application of the combination unit root tests to the post-Bretton Woods US real exchange rate data provides some evidence in favor of the PPP hypo thesis. (C) 2001 Elsevier Science Ltd. All rights reserved.