BUFFER STOCK MONEY AND THE COMPANY SECTOR 1

Abstract

recent contributions) and can now be seen as a viable alternative to the conventional practice of treating the financial balances of the private sector as a single aggregate. There are two strands to the empirical literature on buffer stock money: investigations of how traditional demand for money functions should be modified to take account of financial buffer stocks (examples include Cuthbertson and Taylor (1987), and Muscatelli (1988) for the UK, and Carr and Darby (1981) for the US) and a 'systems' approach which is concerned with both the financial side of the buffer stock hypothesis and also the links between disequilibria in money holdings and other aspects of behaviour (for example, Davidson and Keil (1981), and Davidson (1987)). This paper belongs to the second strand and is novel in that it makes use of cointegration techniques to cut through some of the computational and specification problems associated with earlier work in this area. At the centre of the systems approach is the view that a company's money holdings are determined from moment to moment through its budget constraint by the net volume of transactions in goods, factor services and non-money assets with very little reference to a 'desired' level of money holdings. At the same time companies undoubtedly do have a view as to their desired long run money holdings and over time will endeavour to adjust their money holdings accordingly. Therefore the traditional money demand function should be seen as nothing more than a restriction on the long run behaviour of the company sector. In Section 2 of the paper we use a stylised analytic model to consider how adjustment costs in changing dividends (or any factor which leads to a desire to smooth dividend payments) can generate a role for a financial buffer stock. We show under what circumstances this will lead to past values of this buffer