Raphael Bergoeing: Enrique Mendoza has written an excellent paper. His goal is twofold: first, to quantify the contribution of movements in exchange-rate-adjusted prices of tradable goods and in the price of non-tradable goods relative to tradables as a source of real exchange rate variability, using a thirty-year sample of monthly Mexican data; second, to study whether the observed evidence on exchange rate fluctuations in Mexico is consistent with recently developed theories on sudden stops and the real effects of exchange-rate-based stabilizations in emerging economies. To analyze these issues, he builds a general equilibrium model of endogenous credit constraints with liability dollarization, and he then simulates the model to stress balance sheet effects under large changes in the relative price of non-tradable goods. In particular, he shows that Fisher's deflation mechanism can be an important source of amplification and asymmetry in the response of economies to shocks. The paper also provides key insights into the benefits of inflation targeting rules, a widespread monetary tool in the last decade. This monetary policy, which commits to de jure floating exchange rates, can prove effective in preventing large swings in the real value of the currency.
Traditional models of real exchange rate determination attribute all movements in the real exchange rate to changes in the relative price of nontradable to tradable goods across countries. For instance, Stockman and Tesar, as well as Fernández de Córdoba and Kehoe, develop models in which the real exchange rate is exactly the relative price of nontradables to tradables across countries.1 This setting contains no role whatsoever for movements in the international relative prices of tradable goods. Several recent papers, however, report large and persistent deviations from the law of one price. Most notably, Engel claims that at almost every horizon and for almost every measure and every country relative to the United States, the failure of the law of one price [End Page 136] accounts for over 90 percent (and up to 99 percent) of real exchange rate variations.2Figures 6 and 7 provide alternative support for these mutually inconsistent facts. Figure 6 illustrates the relation between the bilateral real exchange rate for Germany and the United States with a bilateral relative price of nontradable goods. In this figure, rer is the logarithm of the real exchange rate between Germany and the United States, and rerN is the logarithm of the relative price measure. These variables have been constructed so that if the law of one price holds (that is, if the traditional theory works well and if I use appropriated data to measure relative prices), then the two variables should be the same. This figure shows no perceptible relation between the two series, thus supporting the view that the traditional theory is dead. Figure 7, however, provides evidence to the contrary. This figure illustrates the same relation between bilateral variables as in figure 6, but in this case for Canada and the United States. These additional data suggest that the traditional theory is still valid.3
Mendoza's paper demonstrates that figures such as these do not need to be inconsistent. Mexican data support both depending on the exchange rate regime in place. Engel himself claims that a figure such as figure 6 would fairly represent Mexican data. Using a sample of monthly data from 1991 to 1999, he finds that the fraction of the variance of the peso-dollar real exchange rate accounted for by the variance of the Mexico-U.S. ratio for tradables prices adjusted by the nominal exchange rate is over 90 percent, regardless of the time horizon over which the data are differenced. Mendoza shows, however, that Engel may be drawing on too little evidence, as the period he analyzes was characterized mainly by a floating exchange rate regime. Mendoza includes the 1970s and 1980s, when Mexico followed a managed exchange rate regime; he finds that the contribution of the relative price of nontradable to tradable goods becomes much more relevant, accounting for up to 70 percent of the high variability of Mexico's real exchange rate.
This improved characterization of the evidence is...