M0: Causes and Consequences

This paper analyses four aspects of the determination and effects of M0 using annual, quarterly and monthly data. First it analyses the issue of the correct scale variable for determining M0. A combination of survey evidence and statistical tests indicate that retail sales is the most appropriate scale variable to use in looking at M0. Second, the paper looks at the best way to explain the trend in M0 velocity. It confirms earlier results showing that a cumulative interest rate term gives an adequate representation of the trend and is better than a linear time trend for this purpose. Third, the paper looks at the short-run interest elasticity of demand for M0. It finds that this elasticity is highly variable across models and suggests that the analysis of M0 is problematic following substantial changes in interest rates. Lastly, and probably most controvertially it looks at the leading indicators of properties in M0 for inflation. It finds that these are particularly robust and cannot be matched simply by the product of the explanatory terms in the M0 equation. Although this result is implicit in previous work done in the Bank (eg 'VAR models of inflation' Bank of England Quarterly Bulletin, May 1993) this paper makes an explicit empirical link between M0 and inflation.