New Evidence Connecting Exchange Rates to Business Cycles,” Federal Reserve Bank of Richmond Economic Quarterly 84

V irtually every theoretical model of exchange rates predicts that the realexchange rate between two countries (with floating nominal exchangerates) is correlated with the ratio of business-cycle conditions in the two countries. Yet almost no empirical evidence exists to support this prediction of the models. In fact, there is little empirical evidence that ties real exchange rates to any underlying economic conditions. Some well-known studies have concluded that exchange rates appear to have “a life of their own, ” perhaps moving with speculators ’ expectations far more than with changes in economic fundamentals. (See Flood and Rose [1995] for a prominent example.) Contrary to this widely held contention, this article presents new evidence that exchange rates are connected with fundamentals, in particular with the relative gross domestic product (GDP) of each of the two countries involved, as predicted by nearly all exchange-rate theories. Moreover, they are related in the direction predicted by standard models: a country’s currency tends to be depreciated in real terms when that country’s real GDP is relatively high, and vice versa. Why have previous studies not found this relationship between real ex-change rates and ratios of real GDP? The answer is that the relationship is hard to detect with simple linear models, because it appears to be nonlinear and conditional on persistent movements (rather than purely transitory movements) in the data. Recent theoretical and empirical work has pointed to the potentia