This paper examines the performance of the CSO's leading indicators fo
r predicting turning points in the UK growth cycle. Three methods for
extracting a signal are considered: a simple mechanical rule based upo
n the direction of consecutive movements in the leading indicator and
two probabilistic models, namely a standard Bayesian procedure and the
sequential probability model. The results suggest that usefulness of
the shorter leading index is limited and typically outperformed by nai
ve, non-indicator methods. The longer leading index performs only slig
htly better showing a lead time of between four and six months in rela
tion to the reference cycle.