In 2003, Statistics New Zealand received funding under the Government’s growth and innovation framework (GIF) to develop official productivity measures for the New Zealand economy. In March 2006, Statistics NZ published the first official estimates of multifactor productivity for New Zealand. These estimates were for a subset of the industries in New Zealand, referred to as the ‘measured sector’. This measured sector was the original measured sector, which comprised Australian and New Zealand Standard Industrial Classification 1996 (ANZSIC96) divisions A to K, plus P. In 2008, more industries (ANZSIC96 divisions L (pt) and Q) were included in the scope.
Originally, only official productivity estimates were available for the measured sector as an aggregate; no official estimates were presented for individual industries. In March 2010, industry level productivity estimates were released for the first time. Explaining aggregate multifactor productivity growth is important for understanding the performance of an economy.
In principle, a total factor productivity (TFP) residual can be computed for every level of economic activity, from the plant floor to the total economy. These residuals are not independent of each other because, for example, the productivity of a firm reflects the productivity of its component plants. Similarly, industry residuals are related to those of the constituent firms, and productivity in the aggregate economy is determined at the industry level. As a result, productivity at the aggregate level will increase if productivity in each constituent industry rises, or if the market’s share of the high productivity industry increases (and so on, down the aggregation hierarchy). A complete picture of the industrial dynamics of an economy would include a mutually consistent measure of the TFP residuals at each level in the hierarchy and of the linkages used to connect levels (Hulten, 2001).
This paper provides a summary of the work undertaken to date on productivity measurement in New Zealand. It also investigates how user cost of capital is derived, and the resulting methodology that is employed in the current measures. This work is part of Statistics NZ’s current work programme to produce industry level measures of multifactor productivity.
Purpose
Capital (together with labour), is one of the primary inputs in productivity analysis. Statistics New Zealand constructs indexes of industry capital services input as Törnqvist indexes. A Törnqvist index is where the geometric mean of assets’ two-period productive capital stock ratios are weighted exponentially by each asset’s mean two-period share of the industry’s value of capital costs. An asset’s cost of capital is estimated as its user cost (rental price) multiplied by its flow of capital services (proxied by the productive capital stock). The user cost of capital could theoretically be observed as a market rental price for the asset concerned. Given however that capital rental markets are often very sparse (if not non-existent), the user cost is approximated by an implicit rental that owners of capital are inferred to be charging themselves.
The choice of methodology used to derive the user cost of capital services can significantly impact on relative weights assigned to the various asset capital services that provide an input into the production function of an industry. In particular, choosing whether an endogenous or exogenous rate of return is used (or some hybrid thereof), and how capital gains are treated, can markedly impact the resulting asset capital services weights.
Methodology
International recommendation (and theory) gives many alternative methodologies in formulating the user cost of capital, recognising that what suits the unique circumstances of each country will affect choice.
While the preferred method in much of the literature is for an endogenous nominal rate of return with capital gains on assets (excluding land and buildings), a number of factors render the resulting time series of weights in the New Zealand situation economically implausible. This paper presents an analysis of the different methods that could be used to calculate the user cost of capital, and assesses each of the methods against a range of criteria.
Key Results
The use of an exogenous real rate of return (set at 4 percent) and excluding capital gains from the formulation of the user cost of capital provides more economically plausible results in the New Zealand context. The industry asset-weights derived from this method provide a superior representation of the underlying production function of many of the industries. They also display markedly less volatility, and hence are a more realistic representation of the change in the aggregate industry production functions over time. Consequently, this method has been used by Statistics NZ to construct the capital services indexes in the published estimates of multifactor productivity.