The stability of money demand functions: an alternative approach

BIS Economic Papers No 14
April 1985

"No proposition in macroeconomics has received more attention than that there exists, at the level of the aggregate economy, a stable demand for money function."(Laidler, D. (1982), p. 39.)


Few economists would disagree with this view, especially following the adoption by several central banks of aggregate targets for the control and implementation of monetary policy. However, despite intensive analytical and empirical efforts, there is no general consensus concerning the stability (or instability) of money demand functions. This casts some doubt on one of the fundamental assumptions of monetary targeting, viz. the existence of a stable and predictable relationship between money supply and aggregate nominal income.

Sometimes instability is illustrated by unexpected changes in the income velocity of money. More frequently, however, the stability problem is analysed in terms of the money demand function, ie the relationship between money stocks and a few key macroeconomic variables such as aggregate income and interest rates. In order to clarify conceptual differences and to provide a framework for the topics to be discussed in this paper, it may be useful to distinguish between three sources of instability:

  1. Firstly, the income velocity of money will change in response to fluctuations in interest rates as well as to movements in other arguments of the money demand function which are not related to income. Moreover, velocity changes may be observed because of lags in the adjustment of money demand to income. Usually, however, such changes are both predictable and transitory and they can be interpreted as movements along an otherwise stable money demand function with constant lag structures;
  2. Secondly, the money demand function itself may shift, reflecting either unstable parameters or new developments and involving unexpected velocity changes as well. For instance, the process of financial innovation and deregulation may at times accelerate, possibly affecting both the interest elasticity of different monetary aggregates and the balances held at each level of interest rates. Another source of instability may be shifts in the precautionary demand for money, related to changes in confidence, and ongoing institutional changes can create "ratchet" effects so that both current and earlier peak levels of interest rates and aggregate income affect money demand;
  3. Thirdly, since changes in money stocks are induced by movements in money demand as well as by factors on the supply side, the money stocks observed at a given point in time might be "off" the money demand function, unless the speed of adjustment is very high. In other words, over shorter periods the money stocks actually held may not correspond to the money balances desired. Such discrepancies will, of course, induce large and unexpected changes in velocity and can also give the impression that the money demand function has become unstable.

This paper focuses on the analytical and empirical problems related to disequilibrium in money markets and deals only superficially with the many issues that could be discussed under (ii). A distinction is made between a conventional approach which estimates the behaviour of money stocks as if they were demand-determined, and an alternative - or disequilibrium - approach which regards short-term changes in money balances as being determined on the supply side This exogeneity may be explained either by direct actions on the part of monetary authorities or by disturbances in credit markets and in net foreign assets of the banking system. It should be stressed from the outset that in focusing in this paper on the supply of money it is not intended to question the existence of a long-run money demand function as a basis for monetary policy. In fact, one of the purposes of the empirical section is to derive the parameters of the long-run money demand function and to evaluate these in the light of a priori expectations and alternative empirical estimates. The paper does, however, question the notion of a short-run money demand function and proposes an alternative dynamic adjustment path in situations where money markets are not in equilibrium. A whole range of transmission channels could be considered in this respect but only that of interest rate adjustments is being explored. Nonetheless, for several countries this is sufficient to mitigate some shortcomings of the conventional approach, especially the long adjustment lags.

At the same time, since it is not possible to draw a sharp distinction between demand and supply-induced changes in money balances the approach adopted in this paper will also be subject to estimation biases and identification problems. These issues were recognised at an early stage of empirical monetary economics and were usually analysed on the assumption that money markets clear. More recently, empirical techniques have been devised to estimate behavioural parameters in non-clearing markets, using various methods for identifying periods of excess demand and supply respectively. Given the emphasis on analysing the effects of money-market disequilibrium, it would have been natural to apply some of these new techniques. However, like the earlier debate on the identification of money demand and supply functions, the choice of a rigorous and satisfactory estimation procedure could only have been made within a complete model of the financial system, which would be beyond the scope of the present paper.

It should also be noted from the outset that, even though powerful arguments can be found in favour of the notion of buffer stocks and monetary disequilibrium (especially in Laidler (1984), Knoester (1984a and b) and Jonson (1976a and b), there are also strong doubts with respect to the theoretical and empirical relevance of this concept. Kaldor (1980) and White (1981) point out that since money is the most liquid and adjustable asset, a disequilibrium in money markets is unlikely. Moreover, given the easy access to credit facilities, most economic agents would react to real and financial shocks by adjusting their net rather than their gross financial assets. Finally, even though some promising empirical results have been obtained in recent years, doubts also remain in this area, particularly with respect to the robustness and stability of the parameters obtained.

Part I below discusses the main features of traditional money demand equations and of the alternative approaches which pay more attention to supply-side developments. Part II first reviews other studies which have applied the disequilibrium approach either in a complete model or in single equations. It then presents comparable estimates of money demand and interest rate equations for the seven major countries. Part III summarises the empirical results obtained, derives policy implications and considers areas of future work. Annex I is a more technical presentation of the specifications used for the estimations and Annex II discusses alternative lag structures. Annex III gives the data sources and definitions and Annex IV is a selective list of references.