A Synthesis of Keynesian, Monetary, and Portfolio Approaches to Flexible Exchange Rates

The purpose of this paper is to extend our earlier systhesis of monetary and Keynesian approaches to balance-of-payments theory under fixed exchange rates to the analysis of flexible exchange rates. We start with an overview of partial Keynesian, monetary, and portfolio approaches to the determination of exchange rates. We first develop two version of a partial Keynesian model (Meade 1951, Mundell 1962, Fleming 1962) in which the exchange rate is determined at the level that equilibrates the overall balance of payments. We then present three versions of a monetary approach, one assnuming continuous purchasing power parity (Frenkel 1976, Bilson 1978a), and the others assuming either covered interest parity (Dornbusch 1976a) or real interest parity (Frankel 1979). Both versions of the partial monetary approach use domestic and foreign demand-for-money equations as their only other ingredients. Finally, we present a partial portfolio model (Branson 1979) in which the exchange rate and the domestic interest rate vary to achieve equilibrium in a financial portfolio consisting of domestic bonds, foreign bonds, and domestic money. The portfolio approach and the monetary approach are often described as asset market approaches to the exchange rate, because the exchange rate moves in the short run to ensure that existing stocks for assets are willingly held. In the monetary approach, the only assets are national monies, while in the portfolio approach the assets include money and bonds. The partial approaches presented in section I give very different predictions about, for example, the exchange-rate effects of changes in income and interest rates. As a result, strong contrasts have been drawn between them, especially between the Keynesian and monetary approaches. The purpose of this paper, like that of our earlier paper, is to show that most of these contrasts are potentially misleading, since the partial approaches can be best seen as different parts of a larger system. Within the larger system which we develop in section II, we show that the differing predictions of the partial models are based in aprt on ignoring other important parts of the system, and in part on particular assumptions about expectatins and about the strength and speed of the international linkages among national markets for goods and financial assets