The purpose of this paper is to present a simple theory of devaluation for a world economy in which all prices are flexible. The principal theorem of the analysis is that a devaluation and particular changes in the various money stocks have exactly equivalent effects on the current and future equilibrium values of all of the variables of the model. This means that the theory of devaluation is just a particular case of the theory of money. It is assumed that there is no technical progress or changes in tastes over time, that all economic agents have one-period horizons, that all countries in the world have fiat moneys which are the only durable goods traded, that no individual holds the money of any country other than his own, and that there exist exchange authorities which maintain fixed exchange rates. In addition we assume that, except for the changes mentioned in Proposition III.1, each country allows the increments in its stock of money to be entirely determined by currency flows across the exchanges. Under these assumptions there is no loss of generality in working with a two-economy model; and this we do. The devaluing economy will be referred to as Inland and the rest of the world as Outland. Any currency in the latter sector may be used as its numeraire. In addition to Inland's and Outland's fiat moneys, it is assumed that there are n (non-durable) commodities in the world economy, making a total of (n + 2) goods. All commodity excess demand functions are homogeneous of degree zero in money prices, and initial money balances and all demand functions for money are homogeneous of degree one in those variables. We assume also that there is perfect competition and free trade, although the latter of these assumptions is not at all necessary. Finally, we shall allow the model to include both traded and non-traded goods. The model is made dynamic by letting each individual's end-of-period money balance become his initial balance for the next period and by letting his original endowments of commodities be constant over time. We shall assume throughout that this dynamic process is stable, so that beginning with any arbitrary distribution of money stocks in the world ' The first draft of this material was written in 1968, and by now too many