Confidence Intervals for Half-Life Deviations From Purchasing Power Parity

According to the Purchasing Power Parity {(PPP)} theory, real exchange rate fluctuations are mainly caused by transitory shocks. The theory fits well one empirical feature of the data, namely the short-run volatility of real exchange rates, but also implies that shocks should die away in one to two years (the time interval compatible with price and wage stickiness). Existing point estimates of half-life deviations from {PPP} are in the order of 3 to 5 years, too big to be reconciled with the {PPP.} The scope of this paper is to assess how much uncertainty there is around these point estimates. We construct confidence intervals that are robust to high persistence in the presence of small sample sizes. The empirical evidence suggests that the lower bound of the confidence interval is around 4 to 6 quarters for most currencies. With a few exceptions, the results show that the data are not inconsistent with the {PPP} theory, although we cannot provide conclusive evidence in favor of {PPP} either.