Re-examining the Purchasing Power Parity Hypothesis Over Two centuries

This paper reexamines the stationarity of the dollar-sterling real exchange rate using the two centuries of data analyzed in Lothian and Taylor (LT) (1996). We fit univariate time series models for the real exchange rate that dominate the stationary AR(1) specification chosen by LT and the random walk alternative they consider. The differences in our conlusions arise from two primary sources: (1) choice of lag length in the augmented Dickey-Fuller regressions on which the unit root tests are based, and/or (2) realization that the absence of a unit root is necessary but not sufficient for stationarity. Deterministic time trends (and structural breaks) can also give rise to nonstationary RER series. LT (1996) also compare the out-of-sample forecasting results of their stationary AR(1) model with the random work model over the 1973-90 period. They find that the AR(1) forecasts outperform the random walk model. Moreover, the relative superiority of the AR(1) model increases monotonically as the forecast horizon extends from one year to five years. We find that the dollar-sterling real exchange rate is better modeled as either trend stationary process or as a difference stationary process with an MA(5) error. Compared with these two non-stationary specifications, the AR(1) model no longer dominates in out-of-sample forecasts at all time horizons. When the out-of-sample forecasting exercise extends from 1946 onward, our non-stationary models strictly dominate the stationary AR(1) specification. Moreover, the latter forecast exercise indicates the stability of the non-stationary specifications over the pre-float and recent floating rate period. Our results suggest, contrary to LT, that long-run PPP does not hold for the dollar-sterling exchange rate.