Search

The real exchange rate is said to be the single most important price in an economy. While we used to think that we knew what explained its movements (e.g., the Balassa-Samuelson effect), the recent much-cited result by Engel (1999) proposes a serious reinterpretation – i.e., nearly 100% of the movements in the U.S. real exchange rate are explained by deviations from the law of one price. Engel’s finding holds even in the medium run, when movements in the relative price of non-tradables between countries, were thought to be of paramount importance. In this project, we study the movement of real exchange rates based on the prices of Big Macs (which we show are highly correlated with the CPIbased real exchange rates). Our main innovation is to match these prices to the prices of individual ingredients (ground beef, bread, lettuce, labor cost, rent, etc.) in 34 countries during 1990–2002. There are a number of advantages associated with our approach. First, unlike the CPI real exchange rate, we can measure the Big Mac real exchange rate in levels in an economically meaningful way. Second, unlike the CPI real exchange rate for which the attribution to tradable and non-tradable components involves assumptions on the weights and the functional form, we (almost) know the exact composition of a Big Mac, and can estimate the tradable and non-tradable components relatively precisely. Third, we can study the dynamics of the real exchange rate in a setting that is free of the productaggregation bias (argued by Imbs, Mumtaz, Ravn, and Rey, 2002, to be important in studies on CPI real exchange rates), the temporal aggregation bias (argued to be important by Taylor, 2001), or the bias generated by non-compatible consumption baskets across countries. Fourth, we show that Engel's result that deviation from the law of one price is all that matters does not hold generally. Furthermore, deviations from his result can be systematically explained.

The authors exploit three-dimensional panel data on prices for twenty seven traded goods, over eighty eight quarters, across ninety six cities in Japan, and the United States, to answer several questions: 1) Does the average exchange rate between countries stray further from zero, than that between cities within a country? 2) Is there any tendency for the average exchange rate to move closer to zero over time? 3) Does the border narrow over time? 4) Is there evidence linking changes in the so-called border effect - the extra dispersion in prices between cities in different countries, beyond what physical distance could explain - with plausible economic explanations, such as exchange rate variability? The authors present evidence that the intra-national real exchange rates are substantially less volatile than the comparable distribution of international relative prices. They also show that an equally weighted average of commodity-level real exchange rates, tracks the nominal exchange rate well, suggesting strong evidence of sticky prices. Next they turn to economic explanations for the dynamics of the border effect. Focusing on the dispersion of prices between city pairs, they confirm previous findings that crossing national borders adds significantly to price dispersion. Based on their point estimates, crossing the US-Japan border is equivalent to adding between 2.5 and 13 million miles to the cross-country volatility of relative prices. They infer that distance, exchange rates, shipping costs, and relative variability inwages, influence the border effect. After those variables are controlled for, the border effect disappears.