Handbook of Exchange Rates

7.4 Purchasing Power Parity

The literature on PPP is voluminous, and we will not survey it comprehensively here (see instead Rogoff (1996), Taylor and Taylor (2004), Officer (2011)). Rather, we concentrate our discussion on research that can be seen as considering PPP as a generalization of LOP based on the commodity-arbitrage view.

A simple consideration of the enormously high volatility of exchange rates compared to the volatility of relative price levels implies that PPP does not hold continuously. Most of the research on PPP has concentrated on whether PPP holds as a long-run proposition. Much of this has been performed through analysis of the unit root properties of real exchange rates. Early evidence was not encouraging, as studies were unable to reject the null hypothesis that real exchange rates contained a unit root, which implies that the level of the real exchange rate does not return to some stable average level (Mark, 1990).⁹

However, the low power of statistical tests when applied to a relatively short sample span such as the length of the recent float makes rejecting the null difficult. In an illustrative Monte Carlo exercise, Sarno and Taylor (2002) show how, given the observed persistence properties of the real exchange rate, standard unit root tests may be unable to detect stationarity of the real exchange rate (i.e., the validity of long-run PPP), having less than a 50/50 chance to do so for samples spanning up to 100 years. The same problem applies to the LOP ...