RDP 2000-02: Forecasting Australian Economic Activity Using Leading Indicators 2. The Leading Indicators

We examine three leading indices of Australian economic activity: the Westpac-Melbourne Institute Leading Index, the ABS Experimental Composite Leading Indicator and the NATSTAT Leading Indicator (published by the National Institute of Economic and Industry Research). The component series of these leading indices are presented in Table 1. This section discusses the construction methods of each of the leading indices and accordingly, whether we should expect these indices to be useful throughout the cycle, or only at turning points.

The WM index is constructed according to the traditional National Bureau of Economic Research approach whereby component variables are chosen, among other things, on the basis that they represent significant economic processes or are found to be important sources or measures of business cycle movements, that they consistently lead turning points in economic activity and that they conform to general cyclical movements between peaks and troughs (Boehm 1987). The component series of the ABS leading index are chosen on similar grounds (Salou and Kim 1992). In contrast, the component series of the NATSTAT leading indicator are chosen on the basis of the narrower requirement that they consistently and accurately lead cyclical turning points in aggregate economic activity by six months on average (National Institute of Economic and Industrial Research 1993).

Zarnowitz (1992) describes the theory underlying leading indicators as ‘the dynamics of plans and expectations under uncertainty and of institutional and physical constraints in processes of production and investment’. The ABS and WM indices' component series are at least in part chosen on this basis, while the component series of the NATSTAT index are chosen purely for their consistent lead times. Despite the differing criteria for choosing the components of the leading indices, many of the series are similar. All three indices include a measure of labour market pressure, a share price index, a measure of building activity and a measure of business sentiment or conditions. The ABS and NATSTAT indices also include an interest rate series and the ABS and WM indices include a measure of producer prices. The WM leading index contains three extra variables: new telephone installations, M3 and real unit labour costs. The ABS index is the only leading index including a foreign sector variable, US GDP.

The WM index is the only one of the Australian leading indices published in levels – the ABS and NATSTAT leading indices are both published as a deviation from an estimated trend component. Both the ABS and the NIEIR explicitly state that their leading indices are designed to predict turning points in growth cycles in activity, rather than to provide levels forecasts, and this justifies the publication of these indices as deviations from trend rate of growth (National Institute of Economic and Industrial Research 1993; Salou and Kim 1993). Although not as convenient as an index expressed in levels or simple changes, the ABS and NATSTAT leading indices can still be sensibly incorporated into regression models to explain movements in a differenced or detrended activity variable.

In the construction of all three indices, the component series are standardised prior to aggregation to prevent excessively volatile series dominating movements in the index. Once standardised, the series are given equal weight, as weighting schemes typically make no difference to the composite index (Boehm 1987; Salou and Kim 1993). The WM index is graphed in log levels in the top panel of Figure 1 and in log first differences in the bottom panel (scaled by 100). The NATSTAT and ABS leading indices are also plotted in the bottom panel of Figure 1.