Leading Indexes – Do They?

The two recently developed Australian indexes of leading indicators have received much attention in the press. Despite this, relatively little is known about their usefulness in forecasting the associated indexes of coincident indicators (i.e., measures of the business cycle) or any activity variables which move with the business cycle. Using the vector autoregression methodology and the related innovation accounting techniques, this paper evaluates the usefulness of the two leading indexes in forecasting these variables. We also examine the intertemporal relationships between the various variables to determine which (if any) of the variables the leading indexes in fact lead. The results indicate that the two leading indexes are quite useful in forecasting their associated indexes of the business cycle. One of the leading indexes is, however, a lagging indicator of its coincident index. The results for individual activity variables are mixed. The evidence suggests that the leading indexes are useful in forecasting GDP, an index of production and employment related variables. They are not useful in forecasting either retail sales or motor vehicle registrations.