Journal of International Economics 23 (1987) 241-261. North-Holland EXCHANGE-RATE DETERMINATION: AN EQUILIBRIUM APPROACH WITH IMPERFECT CAPITAL SUBSTITUTABILITY Robert DRISKILL and Stephen McCAFFERTY* Ohio State University, Columbus, OH 43210, USA Received July 1985, revised version received November 1986 Two stylized facts of the open economy are often considered prima facie evidence of goods- market disequilibrium: persistent deviations of relative prices from purchasing power parity, and the higher variability of the nominal exchange rate relative to price levels. We develop an equilibrium model characterized by imperfect goods and asset substitutability across countries, and show how it is consistent with the aforementioned stylized facts. We also show that increases in the degree of risk aversion of speculators leads to higher exchange-rate variance if shocks are monetary. I. Introduction International monetary economists now seem to agree on the following empirical regularities or "stylized facts': deviations of relative prices from purchasing power parity are both substantial, persistent, and correlated with deviations of output from trend; exchange rates are more variable than relative price levels. ~ Some economists have taken this as prima facie evidence of disequilibrium, or slow price adjustment, in goods markets but quick or instantaneous adjustment in asset markets. 2 Other economists, though, have recently attempted to explain these stylized facts with equilib- rium models. One class of these equilibrium models, though, can be criticized on a number of points. First, they either fail to explain persistence in deviations of relative prices from trend, or explain them only with the aid of ad hoc assumptions about lags in the supply process. 3 Second, they rely heavily on *We would like to thank Kent Kimbrough and an anonymous referee for comments on an earlier version of this paper. tSee, for example, Kimbrough (1983) and Saidi and Swoboda (1983). 2Perhaps the most provocative presentation of this stance is the well-known empirical work of Frankel (1979). 3Kimbrough (1983) and Weber (1981) have no persistence in their models. Saidi and Swoboda (1983) have persistence, but only because they add a lagged output term to their supply equation. 0022-1996/87/$3.50 © 1987, Elsevier Science Publishers B.V. (North-Holland)