Commentary: On your bike, cyclicals

Those cynics who believe that you can prove anything with statistics will take considerable satisfaction from the behaviour of the cyclical indicators, published by the Central Statistical Office to provide advance warning of peaks and troughs in economic activity.

The indicators - weighted averages of economic, financial and business survey information - agree that the economy reached a turning point in spring this year. Unfortunately, they cannot agree whether it was the start of recovery after a three-year slowdown or the start of a slowdown after an 18-month recovery.

Understandably, the financial markets and the Treasury now virtually ignore the cyclical indicators - and the Central Statistical Office is planning to revamp them by the end of the year.

The cyclical indicators were reasonably good predictors until the 1980s, since when they have fallen victim to the substantial structural changes in the economy and the financial system. A good example is the longer leading indicator (LLI), which is supposed to pinpoint peaks and troughs around 10 months in advance. This forecast a recovery beginning in early 1990 and ending late last year, since when the economy appeared to dip down again.

Of course, no such recovery took place. The LLI was boosted by a rising stock market and unusually over- optimistic responses to the Confederation of British Industry's surveys of manufacturers.

The CSO will have to start again from scratch if the cyclical indicators are ever again to deserve investors' trust.