This article provides a methodological and empirical approach for assessing price level convergence and its relation to purchasing power parity (PPP) using annual price data for seventeen U.S. cities during the period 1918 to 2005. We suggest a new panel data procedure that can handle a wide range of PPP concepts in the presence of multiple structural breaks using all possible pairs of real exchange rates. Testing for PPP requires the definition of parametric restrictions (parity restrictions) across regimes. In general, we find more evidence for stationarity when the parity restriction is not imposed, while imposing parity restriction leads toward the rejection of the panel stationarity. Our results can be embedded in the view of the Balassa-Samuelson approach, but where the slope of the time trend is allowed to change in the long-run.
The Journal of Time Series Econometrics (JTSE) serves as an internationally recognized outlet for important new research in both theoretical and applied classical and Bayesian time series, spatial and panel data econometrics. The scope of the journal includes papers dealing with estimation, testing and other methodological aspects involved in the application of time series and spatial analytic techniques to economic, financial and related data.