The study of turning-points in the business cycle and the construction of leading indicators to provide advance warnings of them is a long-established tradition in economics, but many other economic series also follow cyclical paths in modern developed economies. In Discussion Paper No. 880, Research Fellow Michael Artis, Robin Bladen-Hovell, Denise Osborn, Graham Smith and Wenda Zhang apply the methodology associated with business cycle analysis to UK inflation. They first identify turning-points in the retail price inflation cycle during 1959-90 to show that downswings were consistently shorter than upswings over eight cycles of varying amplitude, so inflation fell much faster than it rose on average. They construct a composite short leading indicator from M0, import unit values and unemployment and a longer leading indicator from vacancies, retail sales, industrial production and world commodity prices. These yield only one `false signal' of a future inflation cycle (although there are more for the individual series).

The authors simulate the use of their indicators to forecast inflation turning-points for both a standard `3CD' method which requires three months of consecutive decline (increase) in the leading indicator before calling the trough (peak) and also a Bayesian approach which also employs information about the distribution of observations within the upswing and downswing regimes. While the Bayesian method improves upon the predictive power of the 3CD approach by sustaining and hence clarifying turning-point signals, the simple method also performs very well. While these results promise improved identification of turning-points in inflation forecasting, a full test can only be conducted in real time in conjunction with conventional forecasts. These results also demonstrate the existence of asymmetry in inflation cycles, despite its absence from the leading indicator series. This suggests that linear time-series modelling is inadequate in this context and highlights the need to develop more appropriate modelling techniques in a non-linear framework.

Predicting Turning Points in the UK Inflation Cycle
Michael J Artis, Robin C Bladen-Hovell, Denise R Osborn, Graham W Smith and Wenda Zhang

Discussion Paper No. 880, January 1994 (IM)