Abstract

Analysis of the Keynesian Liquidity Preference Function

In general, the behaviour of the Keynesian liquidity prefrence function, which embraces an infinitely elastic domain, i. e. the so-called liquidity trap, is either postulated ad hoc or substantiated by reference to a quite specific mode of portfolio behaviour oriented solely to the expected value of the current rate of return per period. It can be shown that this portfolio behaviour, which is generally accepted as constitutive for the shape of the macrorelation, does not legitimate the expected result. This, in turn, not only leads to an incorrect interpretation of the Keynesian demand function for money, but also to false conclusions as to the effect of various monetary policy measures of the central monetary authorities. As long as the number of existing securities is not zero, the equilibrium rate of interest will always lie above the critical minimum rate at which nobody is willing to invest in securities. Making allowance for the overall economic budget restriction, the demand function for money approaches the abscissa asymptotically; an infinitely elastic domain of the function cannot be derived from the underlying behaviour assumptions. Both the demand function for money and the corresponding demand function for securities include the arguments not only of the current and expected rates of interest, but also the amount and structure of the total wealth. Over and above that, the specific behaviour assumptions lead to a pronounced sensitivity of macrorelations to changes in the microstructure, which in particular render approximatively consistent aggregation extremely difficult. If we take account of the wealthdependence of macrorelations and of the mentioned structural characteristics, it is possible to derive a spectrum of monetary policy measures of varying effect, which modifies and extends the traditional Keynesian policy approach in various ways