3.3  The final composite index of leading indicators

The composite index of leading indicators of New Zealand employment was constructed for the period 1985Q2 to 2001Q4 and includes retail sales, new dwelling consents, inventories, short-term overseas arrivals, the spread between the 5 year government bond yield and the 30 day bank bill rate, the Reserve Bank of New Zealand’s trade weighted index (TWI) of the exchange rate, employment intentions, and job advertisements.[6] No foreign activity variable was included in addition to overseas arrivals and the TWI.

All series are released before employment with the exception of inventories. This should make the index particularly useful for forecasting.[7] Inventories are only available up to 2001Q3 and no change is assumed for 2001Q4. Moreover, no job advertisements data are available prior to 1990Q1 and the weights of the component series (and the composite index) were calculated for two periods, 1985Q2 to 1990Q1 and 1990Q2 to 2001Q4. The sets of weights for the component series are reported in Table 1 for the two periods.

Over the first period, 1985Q2 to 1990Q1, inventories, the interest spread and the TWI have the largest weight (0.65 or 0.17 normalised), followed by employment intentions (0.55 or 0.14 normalised). The weights for new dwelling consents and overseas arrivals are 0.5 (or 0.13 normalised), while retail sales with 0.4 or 0.1 normalised have the smallest weight.

Interestingly, for the second period, the weight for retail sales, at 0.72, is almost twice as large as in the first period. This is likely a reflection of the improved economic performance during the 1990s. The increasing weight on retail sales probably also reflects the growing size of employment in the retail sector.[8] The weight on overseas arrivals also increases over the second period from 0.5 to 0.74, probably as foreign trade and overseas economic activity gained importance following the trade liberalisation of the reforms. The normalised weight for retail sales, at 0.13, is higher than in the first period and the only weight that increases when adding employment ads to the index. The normalised weight of overseas arrivals remains at 0.13, while the weight of all other variables is lower (0.12 normalised). Job ads are a particularly good leading indicator and have the largest weight at 0.81 or 0.14 normalised.

Normalised weights are reported in italic.

The composite index of leading indicators is plotted together with employment in Figure 1. The shaded areas in Figure 1 indicate employment downturns from a peak to a trough, where a peak is defined as the quarter prior to at least two consecutive declines in employment and a trough is defined as the quarter prior to at least two consecutive increases in employment.

Employment was hard hit by the economic reforms in the mid-1980s and early 1990s and fell for the six years from 1986 to 1991, apart for 1986Q3 and 1989Q3 to 1990Q3. It rebounded following the end of the recession, in the early 1990s, and grew strongly for several years. From 1996Q4 to 1998Q4, employment once again fell as a result of tight monetary policy in response to inflation pressures from strong growth in the mid-1990s, the negative impact from the Asian financial crisis on foreign demand and two consecutive droughts. Since 1999, employment has been growing steadily apart for two quarters of decline in 2000Q1 and 2000Q2.

Figure 1 - Employment and the composite index

The composite index of leading indicators appears to be effective in signalling cyclical movements in employment. This can be seen from Figure 2, which plots the change in the composite index and employment. The composite index appears to be trending downward prior to a downturn in employment and rebounding before actual employment starts increasing. The relationship between the composite index and employment is evaluated more formally in the next section.

Figure 2 - Change in employment and the composite index