In this paper we consider a particular international economic policy regime: the laissez-faire regime, the distinguishing features of which are unrestricted portfolio choice and floating exchange rates. And as we show, this regime, although favored by many economists, is not economically feasible. It does not have a determinate equilibrium. That is an implication of an overlapping-generations model. More basically, it is an implication of the notion that money is wanted only in order to accomplish trades.

We prove the general existence of steady states with positive consumption in an N goods and fiat money version of the Kiyotaki-Wright model by admitting mixed strategies. We also show that there always exists a steady state in which everyone accepts a least costly-to-store object. In particular, if fiat money is one such object, then there always exists a monetary steady state. We also establish some other properties of steady states and comment on the relationship between steady states and (incentive) feasible allocations.

In "The Inefficiency of Interest-bearing National Debt" (J.P.E. [April 1979]), we argued that private sector transaction costs are needed in order to explain interest on government debt. It follows that if the government's transaction costs do not depend on its portfolio, then, barring special circumstances, an open-market purchase is deflationary and welfare improving. In this paper we show that this result can survive a potentially relevant special circumstances: reserve requirements which limit the size of insured intermediaries.

[In "The Inefficiency of Interest-bearing National Debt" (J.P.E. [April 1979]), we argued that private sector transaction costs are needed in order to explain interest on government debt. It follows that if the government's transaction costs do not depend on its portfolio, then, barring special circumstances, an open-market purchase is deflationary and welfare improving. In this paper we show that this result can survive a potentially relevant special circumstances: reserve requirements which limit the size of insured intermediaries.]

The coexistence of money and default-free interest-bearing government bonds is explained by transaction costs; the private sector absorbs money with less real difficulty than it absorbs bonds. Under the assumption that the costs of issuing money and issuing bonds are identical, it follows that the presence of government bonds is inefficient. Further, the steady-state inflation rate is higher with bond financing of a given real deficit because there is less net output, less real saving, and hence the need for the government to inflate faster. This is demonstrated in a version of Samuelson's pure consumption-loans model.

The coexistence of money and default-free interest-bearing government bonds is explained by transaction costs; the private sector absorbs money with less real difficulty than it absorbs bonds. Under the assumption that the costs of issuing money and issuing bonds are identical, it follows that the presence of government bonds is inefficient. Further, the steady-state inflation rate is higher with bond financing of a given real deficit because there is less net output, less real saving, and hence the need for the government to inflate faster. This is demonstrated in a version of Samuelson's pure consumption-loans model.

[Two competing monetary policy prescriptions are analyzed within the context of overlapping generations models. The real-bills prescription is for unfettered private intermediation or central bank operations designed to produce the effects of such intermediation. The quantity-theory prescription, in contrast, is for restrictions on private intermediation designed to separate "money" from credit. Although our models are consistent with quantity-theory predictions about money supply and price-level behavior under these two policy prescriptions, the models imply that the quantity-theory prescription is not Pareto optimal and the real-bills prescription is.]