An examination of the monthly Swiss franc/U.S. dollar rate over the 1954–86 period indicates that nominal and real exchange rates follow a random walk. That is, changes occur unexpectedly but are anticipated to be permanent. Exchange rate changes are thus not useful for predicting currency price movements. Furthermore, variations in nominal exchange rates are dominated by changes in real rates. Past rate changes will not be useful for predicting inflation, or vice versa.

Changes in money supplies, real incomes, interest rates and current or trade accounts appear not to be systematically related to changes in exchange rates. In the absence of any systematic component that could be exploited for forecasting purposes, what means are available for controlling exchange rate risk? Hedging can be used with some effect, but the optimal hedging policy is still a matter of debate. Given the relatively low correlations between Euromarket interest rates and returns on common stock indexes and relative changes in exchange rates, diversification across international markets offers considerable potential for risk reduction.