4.2.2 Estimates of multifactor productivity by industry
Industry multifactor productivity estimates were constructed by dividing the chained Laspeyres volumes GDP index by a Fisher input index. The Fisher input index was formed by combining industry hours worked and industry capital stock data, using industry labour cost and industry capital cost data as weights. Industry multifactor productivity series are presented in Appendix 2.
| March Year | Primary | Mining and Quarrying | Construction | Manufacturing | Electricity, gas and water | Transport and communications | Business and property services | Personal and community services | Retail and wholesale trade |
|---|---|---|---|---|---|---|---|---|---|
| 1988 to 1993 | -0.52% | -1.91% | -4.59% | 0.29% | 1.11% | 6.75% | -2.54% | 0.82% | -0.38% |
| 1993 to 2002 | 2.45% | 0.72% | 0.25% | -0.16% | -0.93% | 5.52% | 0.74% | 1.48% | 1.40% |
| 1988 to 2002 | 1.38% | -0.23% | -1.51% | 0.00% | -0.21% | 5.96% | -0.44% | 1.24% | 0.76% |
Over the 1988 to 2002 period, the productivity growth of two industries stands out: the Transport, storage and communications industry (average growth of 6.0% per annum) and the Primary industry (average growth of 1.38% per annum) (see Table 4 and Figures 5 and 7). The Personal and community services industry also appears to have had comparatively strong average productivity growth (1.24% per annum, see Table 4 and Figure 6). However, for the reasons discussed above, we view this result with less confidence.
At the aggregate level, New Zealand’s productivity growth appears to have accelerated post 1993. Of the nine industries presented here, estimates for six industries show a stronger productivity performance in the period 1993 to 2002 than for the period 1988 to 1993 period. These industries are the Primary, Retail and wholesale trade, Business and property services, Mining and quarrying, Construction and Personal and community services. For all these industries, the increase in average productivity growth between the 1988 and 1993 and 1993 and 2002 was in excess of 0.6% per annum.
The three industries that did not experience an acceleration in productivity growth post 1993 are the Electricity, gas and water, Transport and communications and the Manufacturing industries. Multifactor productivity in the Transport and communication industry was increasing relatively strongly in the period 1988 to 1993 (6.75% per annum), and the rate of productivity growth slowed slightly in the period 1993 to 2002 (5.52% per annum). The Electricity, gas and water industry, which experienced a modest increase in productivity between 1988 and 1993 (1.11% per annum), appears to have declined sharply since the mid 1990s (see Figure 6), resulting in a decline in productivity over the period 1993 to 2002. The manufacturing industry, having experienced some positive productivity growth in the period prior to 1993, experienced a slight decline post 1993. Estimates suggest that manufacturing industry productivity in 2002 was very similar to productivity in 1988.
Some of these differences in productivity performance across industry and across time periods are likely to reflect the consequences of industry reforms. Sub-industries within the Transport and communication industry, such as telecommunications and postal services, along with the Primary industry, were some of the first to undergo reform in the mid 1980. The benefits of these reforms, and the benefits of later reform in other industries, may be reflected in these results. However, results for the manufacturing industry, which has experienced considerable reform, show no multifactor productivity growth. Manufacturing industry productivity results also show surprisingly little volatility. These differences are in productivity performance are consistent with Buckle, Haugh and Thomson (2001) who found a decline in volatility for manufacturing GDP growth, but also that manufacturing sector GDP growth was slower than for the Primary or Services sectors.
Section 4.2.1 noted the possible impact of omitted variables in biasing upwards results for individual industries. Land, one of the key inputs to the primary industry, is omitted from calculations. The extent to which changes in land use patterns and land prices have impacted upon results for the Primary industry are unknown. The Primary industry also includes fisheries. The introduction of fisheries quotas will have increased firm costs and therefore decreased value added in this industry. However this is only a partial consideration of resource value, as the resource input is not included in the input index.
Because measuring service sector output is difficult, multifactor productivity growth for the three service industries should be viewed with caution. Retail and wholesale trade, which spans both the goods and services sectors, has been the third strongest performing industry in terms of productivity growth since 1993. This industry appears to have experienced a fall in multifactor productivity during the early 1990s, which coincided with a decline in New Zealand’s domestic absorption rate.
These results show considerable variation in the productivity growth between industries. While some sectors have shown notable growth (Transport and communications, Primary), others have stagnated. When looking at these results, it must be remembered that output of some industries, particularly those in the service sector can be difficult to measure. These measurement problems are recognised by the Australian Bureau of Statistics, and consequently their productivity measures exclude some parts of the service sector.
4.2.3 Comparison with Diewert and Lawrence (1999)
Diewert and Lawrence’s Official Database included series for 20 industries in the market sector of the New Zealand economy, for the period 1978 to 1998. Industry analysis in this paper has been at a considerably more aggregated level, looking at nine industries, due to the data limitations discussed in section 3.
Diewert and Lawrence reported their industry level results using trend growth rates. In order to create comparability with the Diewert and Lawrence industry results, we have calculated geometric compound growth rates, our preferred measure, for their industry multifactor productivity estimates over the periods 1987 to 1998 and 1993 to 1998.
Because results for Diewert and Lawrence’s industries have been re-calculated using exactly the same index number and growth rate calculation formula, the differences in the results reflect underlying differences in the databases. These are considerable, including the change from using the Philpot capital series to the Statistics New Zealand capital series. Reflecting this, the two studies share some similarities, but notable differences, in their results.
Like the multifactor productivity results, the Diewert and Lawrence official database also suggests that Transport, storage and communications and Primary have been the industries with fastest growing productivity. In addition, the magnitudes of the growth rates appear roughly comparable.
Over all industries, the Diewert and Lawrence results appear to be more positive over the 1987 to 1998 period. Contrasting this, multifactor productivity results appear more positive in the shorter period from 1993. More positive results in the shorter period are a pattern across the three service industries. However, the multifactor productivity growth in Business and Property services industry is more positive across the entire 1987-1998 period. These results may reflect ongoing measurement improvements in what are hard to measure sectors, particularly the methodological change, discussed above, in the business and property services industry.
Multifactor productivity results suggest that the Electricity, gas and water industry experienced poor growth in productivity. Diewert and Lawrence also reported on this industry and their dataset produces a considerably higher estimate of productivity growth. Manufacturing multifactor productivity also performed poorly and this is similar to the results for manufacturing industries reported by Diewert and Lawrence (1999).
