The Timing of Payments and the Demand for Money

Abstract

Money is the asset which can be used to make payments. The classical theory of the demand to hold money balances rests on the lack of synchronisation between the time patterns of receipts and of payments. Because, for any individual, every payment is not matched by a simultaneous and equal receipt, some money is held in anticipation of future expenditures. The classical analysis assumed, moreover, that the time patterns of receipts and of payments are only capable of slow change over time. That is, the payments habits of people and institutions were regarded as lacking in flexibility in the short run. This assumption, which is still generally accepted,2 underlies the Keynesian postulate that the transactions demand for money is a stable increasing function of national income. The ratio of the money stock to national income (and therefore its inverse the income velocity of circulation) has varied considerably over time. This fact, together with an acceptance of the classical assumption of inflexible payments habits, led Keynes and others to introduce variables, additional to national income, into the demand function for money. Keynes expressed this extension in terms of there being other motives (the precautionary and the speculative) as well as the transactions motive for holding money. This form of language, together with the notational distinction between M1 and M2 balances, has been responsible for the widespread practice of dividing the demand for money into two parts, the one related to the making of payments and the other to the choice between assets to hold. The latter part has been at the centre of post-Keynesian theorising about money; but as regards the former, there is general agreement with Keynes' acceptance of the classical assumption. A theory of the timing, as distinct from the volume, of payments is worked out in this article. It shows that the time-lags between payments are variable, and therefore rebuts the classical assumption. The analysis has three aspects: (i) That changes in the expected yield rates on real assets can alter the timing of transactions, especially those in financial assets (including trade credit), and thereby alter the demand for money. I I am indebted, for valuable comments, to L. R. Klein and P. Davidson of the University of Pennsylvania and to W. J. Baumol of Princeton University. ' A notable exception in this respect is R. Turvey. See his Interest Rates and Asset Prices, 1960, chap. 3, and On the Transactions Demand for Money , in 25 Economic Essays in Honor of Erik Lindahl, 1956.