Working Paper, Department of Economics, Queen Mary, University of London 484
The empirical literature that tests for purchasing power parity (PPP) by focusing on the stationarity of real exchange rates has so far provided, at best, mixed results. The yen real exchange rate behavior, as compared to other major currencies, has most stubornly challenged the PPP hypothesis and deepened this puzzle. This paper contributes to this discussion by providing new evidence on the stationarity of bilateral yen real exchange rates. We employ a non-linear version of the Augmented Dickey-Fuller test, based on an exponentially smooth-transition autogregressive model (ESTAR) that enhances the power of the tests against mean-reverting nonlinear alternative hypotheses. Our results suggest that the bilateral yen real exchange rates against the other G7 and Asian currencies were mean reverting during the post-Bretton Woods era. Thus, the real yen behavior may not be so different after all but simply perceived to be so due to the use of a restrictive alternative hypothesis in previous tests.
PPP, Yen, Real exchange rates, Nonlinear models, ESTAR models