Working Paper, Federal Reserve Bank of Atlanta 99-21
The half-life of deviations from purchasing power parity (PPP) plays a central role in the ongoing debate about the ability of macroeconomic models to account for the time series behavior of the real exchange rate. The main contribution of this paper is a general framework in which alternative priors for the half-life of deviations from PPP can be examined. We show how to incorporate formally the prior views of economists about the half-life. In our empirical analysis we provide two examples of such priors. One example is a consensus prior consistent with widely held views among economists with a professional interest in the PPP debate. The other example is a relatively diffuse prior designed to capture a large degree of uncertainty about the half-life. Our methodology allows us to make explicit probability statements about the half-life and to assess the likelihood that the half-life exceeds a given number of years, without taking a stand on whether or not the data have a unit root. We find only very limited support for the common view in the PPP literature that the half-life is between three and five years.
Econometric models Purchasing power parity Time-series analysis