5 Data issues that influence productivity measurement and research
This section addresses the measurement of inputs and outputs that may influence measured productivity. First the measurement of two of the most important inputs into the production process, namely capital and labour, is considered. Then the measurement of the quantity of output is discussed.
5.1 Capital
The flow of capital services is not usually directly observable and most studies assume that the value of capital services is proportional to the value of the capital stock. There are two broad approaches to measuring the capital stock; these are measures of the gross capital stock and measures of the net capital stock. The gross capital stock approach assumes that a capital good delivers a constant flow of services over its lifetime. Constructing a gross capital stock measure involves summing past purchases of capital goods (ie, investment) in constant prices back as far as the assumed life of the asset. Different types of capital goods will have different assumed asset lives, for example a building will have a longer life than a van. Equation expresses the gross capital stock approach:
(20)
Net capital stock measures allow the service flow from an asset to fall over time as the asset deteriorates. That is, net capital stock measures allow for depreciation of an asset. As shown in equation , the capital stock in time t is equal to the previous periods net capital stock less depreciation plus any additional capital investment made:
(21)
where δ is the depreciation rate, or proportion of the capital stock that is retired each period. To construct a net capital stock series, a starting value for the net capital stock is required.
Whether to use gross or net capital stocks in productivity studies has been the subject of some debate. For example, Philpott (1992) stated “For use in analyses of production and productivity (for which we intend using our own series) net capital measures are less appropriate than gross capital measures.” Diewert and Lawrence (1999) state:
Some argue that the gross capital approach of effectively assuming no deterioration is more appropriate for Buildings and structures while the net capital stock approach of assuming a steady deterioration is more appropriate for Plant and equipment. … [T]he ABS actually takes weighted averages of the gross and net capital stock estimates with the weights varying by asset type to more closely approximate the deterioration patterns estimated by the US Bureau of Economic Analysis. In general, we believe the net capital stock approach provides a closer approximation to the contribution of capital assets to the production process and better approximates economic depreciation.
Diewert and Lawrence, 1999: 40
Up until recently there have been no official estimates of the capital stock for New Zealand and consequently researchers have needed to form their own estimates. Any estimate of the capital stock necessarily has to make a number of assumptions, including assumptions about the asset lives of the different capital assets under consideration; the appropriate depreciation rates (when using a net capital stock concept); and how to aggregate different types of capital assets to form an overall measure of the capital stock. Consequently, Diewert and Lawrence (1999: 39) note that “Even in countries where official capital stock estimates are available, many researchers prefer to form their own estimates as their beliefs about the economic characteristics of capital may differ from statistical agency practice.”
Prior to the recent release of an official capital stock series by Statistics New Zealand, the work of Philpott (1992) provided a useful source of annual gross capital stock data. Philpott (1992) provided estimates of the stocks of two types of capital asset types (Building and Construction, and Plant and Equipment) for 22 SNA production groups for the period 1950 to 1990. The estimates were based on the “Perpetual Inventory Method” (PIM).
Since Diewert and Lawrence (1999) published their Treasury Working Paper Measuring New Zealand’s Productivity, Statistics New Zealand has constructed an official capital stock series back as far as 1971/72. Data on the “productive” capital stock are available, which represents accumulated investment less the accumulated value of the assets retired and the loss of efficiency of those assets still operating. The official capital stock series have been developed using a perpetual inventory model and provide stock data for a number of different asset classifications. The different asset types are: residential buildings, non-residential buildings, other construction, transport equipment, plant and machinery and the intangible fixed assets of mineral exploitation and computer software. Capital stock data are also available at the industry level.
Once capital stock measures for various asset types are available, the researcher then needs to assign each of the asset types a user cost in order to form a total input index. User costs of capital are the implicit rent that the owners of capital goods ‘pay’ themselves in return for the goods services. Measuring user costs is a significant task and the reader is referred to section 5.4 of OECD (2001) and/or appendix D of Diewert and Lawrence (1999) for more details.
