|
Many argue that `financial innovation' destabilized money demand in
the 1980s, which undermined policies based on targeting monetary
aggregates and thus led governments to adopt alternative targets, such
as money GDP or the exchange rate. The instability of the broader
aggregates as found in the UK in the 1980s need not apply, however, to
all measures of the money supply. In Discussion Paper No. 735, Research
Fellow Michael Artis, Robin Bladen-Hovell and Dilip
Nachane argue that the traditional, direct approach of identifying a
money demand function and using conventional tests to look for
structural breaks runs into measurement errors and cannot disentangle
the effects of supply and demand. They focus instead on the `velocity'
of money (the ratio of nominal income to money) as the sole measure of
its demand and test for instability using a method drawn from spectral
analysis. By decomposing time-series into an `evolutionary spectrum' of
cyclical, `long-run' and noisy, `short-run' components, this method
looks for structural breaks by testing for shifts in the weights of
these components over time. It also avoids the very strong assumptions
entailed by common methods of identifying structural breaks, in
particular those concerning the underlying distribution of the series. |