Stats › Currencies and commodities › PPP
Purchasing power parity
Purchasing power parity is the idea that a basket of goods should cost the same everywhere once you convert at the exchange rate. If it doesn't, the exchange rate is supposed to move until it does.
It is the most intuitive theory in international finance and it is close to useless over any horizon a trader cares about. The consensus estimate is that deviations from PPP take three to five years to halve. Some estimates put the upper bound of the confidence interval past thirty years. And the single most cited result in the whole field, Meese and Rogoff in 1983, is that a coin flip beats the structural models out of sample.
I put this page together because PPP gets used as a valuation anchor in a lot of retail commentary, usually via the Big Mac index, and the literature on how slowly it works almost never comes with it. So here is the literature, with the numbers, the samples and the people who measured them.
Rogoff's 1996 survey states that "the consensus among these studies on the half-life of PPP deviations is remarkable (three to five years)". Individual estimates he cites run from 2.8 years to 4.7 years. Murray and Papell, using median-unbiased estimation on the dollar-sterling rate from 1791 to 1990, get a point estimate of 6.58 years with a 95% confidence interval of 3.72 to 34.31 years, and infinite upper bounds under some specifications. Meese and Rogoff, forecasting from Nov 1976 over a Mar 1973 to Jun 1981 sample, found a random walk performed as well as any estimated model at one to twelve month horizons even when the structural models were fed the realised values of their own explanatory variables. On The Economist's Big Mac index the Swiss franc was about 45% overvalued and the cheapest burgers about 61% undervalued in the edition published 30 Jul 2026.
Members only
The rest of this page is for members
Below this point there are 8 sections, 1 chart, 1 table and 9 named sources, roughly 2600 words of it. Every figure carries the source it came from and the date the data is from.
You can keep browsing every statistic in the library for free. The intro and the summary are always open.