Macro-Economic Analysis of the Demand for Money

Summary

The factors which would determine the demand for money at the micro, individual level — it was argued in chapter 3 — were largely expectational considerations of the future level and variance of cash flows, of asset prices, etc. It has not been easy to find quantitative counterparts, at least at the macro-economic level, of such behavioural factors. The tendency has been to turn instead for the study of behaviour at the aggregate level to a rather more generalised, simpler theory of asset demand, relating the demand for each asset to total wealth (as the budget constraint), relative interest rates and other variables, such as the level of incomes, which affect the demand for the specific services of the various assets. This gives a demand for money function of the form:

$${M_d} = f(Y,{r_i},W)$$

However, wealth is the capitalisation of present and future incomes, so the above explanatory variables are not fully independent. The normal procedure is to omit either W or Y from the estimating equation. Much of the controversy between ‘monetarists’ and ‘Keynesians’ (or ‘fiscalists’) has focused on the particular form of the demand for money function, whether it was stable and predictable, and whether the interest elasticity of demand for money (determining the response of holders of money balances to changes in yields on alternative liquid financial assets) was high or low. The question of the (real) income elasticity of demand for money, though of interest in itself, has engendered less heat, since no crucial point of theory rests on this issue. Monetarists took the view that the demand for money function is relatively stable and predictable, and the interest elasticity of demand for money relatively low (at least in comparison with the interest elasticity of demand for goods); Keynesians argued the opposite.

The results of putting these hypotheses to empirical test are described in Section 2. The question of the interest elasticity of the demand for money has been satisfactorily resolved with the finding of significant but quite low values. Econometric results and views concerning the stability of the demand for-money function have, however, changed over the years as the passage of time generated new data. Earlier studies prior to 1973 suggested stability, but during the disturbed years 1973–6, such demand-for-money functions predicted quite poorly. This failure was attributed to financial innovation and temporary supply shocks. Nevertheless, the more general relationships between accelerations and decelerations in monetary growth and in nominal incomes held up well and most monetary authorities adopted monetary targetting. After 1979, however, the combination of a switch in policy with further supply shocks led to more serious breakdowns in the stability of the relationships between money, incomes and interest rates. Previous single-equation estimates of the demand-for-money have not been able to take adequate account of these supply shocks, policy regime changes or more fundamental structural changes. Such changes are breaking down the functional divisions between financial intermediaries and may even be putting into greater question the clarity of definition of ‘money’ itself.