4.2  Industry productivity

The market sector productivity series provide an understanding of aggregate trends in New Zealand’s productivity growth. However, aggregate trends can mask differences in industry productivity trends. Buckle, Haugh and Thomson (2001) found considerable changes in industry growth rates and the proportion of GDP that these industries comprised. In particular, the primary and services sectors showed increasing growth rates. Multifactor productivity results suggest that changes in GDP growth rates across industries may reflect changes in aggregate productivity growth. This section examines multifactor productivity trends in the nine industries comprising the market sector of the New Zealand economy.

4.2.1  Industry level measurement issues

Industry based multifactor productivity series should be approached with caution due to considerable measurement problems at this level of disaggregation. While many of the problems discussed here have implications for results at the aggregate level, these are likely to have a more significant impact on industry level results. For example, the most basic problem – a lower rate of surveying of the population at an industry level – is overcome through aggregation. In contrast, the lower survey rate at the industry level is likely to result in greater volatility in the industry series.

The key measurement issues raised by Diewert and Lawrence (1999) included double deflation in GDP, omitted variables, misallocation of demand, and measurement of financial sector output. Problems with capital stock data and consistency of definition of industry between labour data series were also raised. These last two issues were discussed in section 3 of this paper.

Double deflation is the ideal method of calculating industry based GDP. This requires price and quantity information on gross outputs and intermediate consumption purchases between industries for every period. In practice, sufficient information for each industry to calculate GDP on a double deflation basis is not readily available, and is therefore only used in a few industries. A comprehensive source of information for double deflation is input-output tables, but these are generally only produced on a 5-year basis, with significant lag times between collection and analysis. Statistics New Zealand has introduced double deflation to a further industry, business and property services, since Diewert and Lawrence’s (1999) report (Statistics New Zealand 2002). The alternative method of calculating GDP is the single indicator method. This method assumes that the structure between industries is constant, although the weighting assigned to each industry is updated annually.

The calculation of multifactor productivity would ideally take into account a number of currently omitted variables such as land use, natural resource depletion (eg, fishery and forestry stocks) and inventories.[10] Measures of these items are not currently available in New Zealand. Changes in resource use and prices would lead to changes in multifactor productivity growth. At the industry level, differences in usage of land, natural resources and other unaccounted inputs between sectors may create misleading results. For example, land, a key input into agricultural production, has not been valued, while the key inputs in other industries may have been more comprehensively considered (eg, capital services in the manufacturing industry).

Misallocation of demand can arise when intermediate industry consumption is not correctly identified, or allocated to the wrong industry. For example, consider the situation when a firm provides an employee with a company vehicle. The cost of the vehicle is an expense to the firm, reducing GDP. By contrast if the firm paid the worker more in order to purchase their own vehicle, this would increase final demand and GDP. Changes in policies leading to altered incentives for providing fringe benefits, or leading to more self-employed people, may lead to changes in the way consumption is split between intermediate and final demand.

By its nature, output from the service sector can be extremely difficult to measure. This is particularly so in the financial services industry because output depends on net interest earnings. These must be split between intermediate or final consumption and between the different industries of the economy. Value added in other service sector industries, such as education, can also be hard to ascertain because of the difficulty in identifying a measurable output. In some cases employment indicators may be used along with an assumed constant employment to output ratio. In other words, for some of the sub-industries within the service sector, constant labour productivity may be assumed (Statistics New Zealand 1996). Difficulties in measuring services sector output has led to the exclusion of parts of this sector from the Australian Bureau of Statistics multifactor productivity series (see Section 5).

Overall, measurement issues at the industry level suggest industry multifactor productivity series should be treated with a degree of caution. This is particularly the case for service sector industries where measurement problems are more pronounced.