Another look at long-run purchasing power parity using Sims tests for unit roots

Abstract

This paper tests for long-run PPP using two types of unit root tests, standard Dickey-Fuller (DF) or Augmented Dickey-Fuller (ADF) tests and Bayesian Sims tests. A problem with DF and ADF tests has been the low power of the tests; that is, such tests have difficulty in rejecting the random-walk hypothesis for real exchange rates. Sims tests, in contrast, do not give undue weight to the unit root. Instead Sims tests use a Bayesian posterior odds ratio which spreads the probability uniformly on the [0,1] interval. Monthly data for several industrialized and less-developed countries are used. Results indicate that real exchange rates do not appear to follow a random walk. While short-run deviations from PPP do occur, they fade in the long run.