6 Conclusions
This paper has provided new market sector and industry productivity series for the New Zealand economy for the period 1988 to 2002. These series were constructed using an industry database containing official Statistics New Zealand data, including upgraded National Accounts and productive capital stock data that were first released in 2000. Throughout, productivity series have been constructed using index number techniques, building on the substantial work of Diewert and Lawrence (1999) in Measuring New Zealand’s Productivity.
Measured over alternative business cycles, average multifactor productivity growth for the market sector of the New Zealand economy ranged from around 0.8% to 1.2% per annum. Average labour productivity growth varied in a much tighter range between approximately 0.7% and 0.9% per annum. There appears to have been a noticeable improvement in market sector multifactor productivity after 1993. Average multifactor productivity growth increased from 0.09% per annum in the period 1988 to 1993 to 1.32% per annum in the period 1993 to 2002. This result is consistent with both a structural improvement in New Zealand productivity growth and earlier research showing improvements in New Zealand GDP growth dating from the early 1990s (Razzak, 2002; Buckle, Haugh and Thomson, 2002). However, owing to the short period covered by the productivity data it was not possible to use formal tests for structural breaks in New Zealand’s productivity after 1993.
The comparable Diewert and Lawrence (1999) productivity series to those reported in this paper were constructed using the ‘Official’ Database. In general the New Zealand market sector productivity series were similar to the Diewert and Lawrence (1999) productivity series, although the two capital productivity series appeared to diverge from the mid-1990s.
Multifactor productivity growth has been strongest in the Transport and communications industry, followed by the Primary industry. Diewert and Lawrence (1999) also reported strong multifactor productivity growth in these industries. Excluding hard to measure industries from the market sector, to form ‘ABS equivalent’ productivity series for New Zealand, resulted in significantly higher average multifactor and labour productivity growth in New Zealand.
In the period up to 1993, labour productivity growth was higher in New Zealand than in Australia, however between 1993 and 2002, labour productivity growth was lower in New Zealand. Because multifactor productivity growth has been similar in both countries, the difference in the evolution of capital-labour ratios in the two countries accounts for the difference in labour productivity growth. The impact of changes in labour market regulation and welfare reform on New Zealand’s capital accumulation and labour productivity growth warrants more detailed investigation.
This paper has provided a basis for further work on New Zealand’s productivity. One strand of work is to improve the coverage and quality of the industry database used in this paper. Existing data could be backdated further and additional inputs (such as human capital, land and inventories) could be collected or constructed. This could also include the construction of industry capital stock series that have been weighted using user cost of capital series to aggregate asset type capital stock data at the industry level. Sensitivity analysis of the productivity series to alternative data (for example, the QES hours paid data) could also be undertaken. Another strand of work is to use the industry productivity series to examine the industry sources of aggregate productivity growth. This will provide policy makers with a basis for evaluating why some industries have contributed more to aggregate productivity growth than have others. Another area that warrants investigation is the relative importance of technological change versus economies of scale in generating industry productivity growth, using a technique proposed by Diewert and Fox (2003). Insights from this work are important as policy settings are likely to differ depending on whether scale economies or technological change is the main driver of productivity growth.
