The US Conference Board has overhauled its leading indicator index, making changes to four of the 10 components. The new version has performed more weakly in recent years and over the last 12 months – see chart. It has recovered, however, from a low in September to reach a five-month high in December.

The revision is of questionable value. The Conference Board claims that the new index is a more accurate predictor of business cycles since 1990 but its earlier performance is inferior – as the chart shows, it failed to turn down before the 1960-61 recession, in contrast to its predecessor.

An important change is the replacement since 1990 of the real M2 money supply with a new “leading credit index”. This partly explains why the new index weakened last summer while the old version continued to rise. This weakness, however, appears to have been a false signal, based on recent solid US economic data. Had the new index been in operation, in other words, it would have encouraged dubious recession calls.

The view of this journal, of course, is that real money is a key forecasting tool and should be included in a composite leading index, although in most countries a narrow measure outperforms M2 and broader aggregates.

Recession-mongers will probably claim that the new index is consistent with their forecast since, despite the recent recovery, it has yet to regain the July 2011 peak. They can also argue, with some justification, that the index remains biased upwards by its inclusion of the 10-year Treasury yield / federal funds rate spread – the assumed positive implication of a steep yield curve is questionable when official interest rates are close to zero*. (The view here is that this bias is counterbalanced currently by a downward distortion from the omission of real money.)

* The economy entered a recession in 1937 with a similar Treasury yield curve to currently.