Panel stationarity tests for purchasing power parity with cross-sectional dependence

We investigate the purchasing power parity (PPP) hypothesis for a group of 17 countries using a new panel-based test of stationarity that allows for arbitrary cross-sectional dependence. We treat the short-run time series dynamics nonparametrically and thus avoid the need to fit separate, and potentially misspecilied, models for the individual series. The statistic is simple to compute and uses standard normal critical values, even in the presence of a wide range of different deterministic components. We also evaluate the behavior of the test using a factor model to approximate cross-sectional dependence and find that it generally improves finite-sample performance. Taken together, these features provide a widely applicable solution to the problem of testing for stationarity versus unit roots in macro-panel data. The test finds significant evidence against the PPP null hypothesis being true, even when allowance is made for structural breaks.