The econometric estimation of the demand for money

Despite thirty years of research there is still widespread disagreement about even the basic explanatory variables which should enter into a demand for money function. Two reasons for this stand out, the dual role of money as transactions and savings medium, and the many different definitions of `money'. This thesis aims to cast some light on the demand for money by conducting work using three different approaches and on three different aggregates. The work is grounded in surveys of both the economic and econometric literature. The first piece of research looks at MO and uses the General to specific methodology developed by Hendry et al to derive a final equation which we believe to be the best representation of the data available. The importance of rigorous testing is stressed within this framework, and our analysis lays stress also upon the need for thorough economic analysis at all points in the research. The second research looks at M1 and is essentially an applied examination of the role of risk terms in the demand for money. A theoretical portfolio model is set up, and econometric analysis is applied to test both the overall and individual significance of the risk terms which the analysis suggests could be of importance. Finally the research considers broad money, and examines how buffer stock theories can be applied to sectoral behaviour. The emphasis is on the consideration of variables which such theories would lead a researcher to consider, rather than on developing the detailed micro-theoretic background. (DX89572)