Real exchange rates are quite persistent. Standard unit root tests are not very powerful in drawing a conclusion regarding the validity of purchasing power parity [PPP]. Rather than asking if PPP holds throughout the whole sample period, we examine if PPP holds sometimes by employing Hamilton-type (1989) Markov regime switching models. There are various reasons that the persistence of real exchange rates changes over time. When at least one of multiple regimes is stationary, PPP holds locally within the regime. Employing 5 real exchange rates spanning more than 100 years, we find strong evidence that the strength of PPP is changing over time. We make comparisons to an early work throughout the article. The new model selection criterion, provided by Smith et al. (2006), called the Markov switching criterion devised especially for discriminating Markov regime switching models, unambiguously indicates a preference for the Hamiltontype Markov regime switching model employed in this article. Also, the evidence for PPP is not much different during the Bretton-Woods and current float periods whether PPP holds or not.
Regime switching real exchange rates Markov switching criterion purchasing power parity