|
This research will analyze the information technology productive efficiency and productivity of some industries of Australia. We employed stochastic production functions instead of the traditionally used deterministic functions. The benefits of using stochastic functions are as follows. First, it recognizes the fact that companies or industries do not always produce at their optimal capacity. Second, it allows us to measure technical efficiency and productivity while non-stochastic approach only allows us the latter.
Our analysis is also under the profit maximization assumption. That is, firms can decide the input and output quantities based on the market information, i.e., input and output price changes. Thus, the quantities are endogenous rather than exogenous. With this assumption, we not only can generate unbiased and consistent estimations, but also can estimate allocative and scale efficiency.
An IT productivity model may be expressed as a linear (or log-linear) relationship between output (e.g., revenues) and input (e.g., costs). The parameters of the model may be estimated by using the multivariate regression technique. The results of a multivariate regression should be interpreted as correlational (and not as causal) because independent variables (the right-hand side) in a multivariate regression model are considered exogenous, and uncorrelated to each other. This translates to "the firm does not have any control on the input quantities," and hence, input quantities vary randomly. However, there is dependence between input variables, which arises because managers attempt to maximize profits or minimize costs by making a simultaneous choice in input quantities. Statistically speaking, the parameters estimates of the production function will be inconsistent if the dependence is not addressed in the empirical model.
Three types of inefficiency, technical, allocative, and scale inefficiency can be measured from our methodology. Technical inefficiency measures the difference of the actual output from the optimal level. Allocative inefficiency measures the difference of the input allocation between the actual level and the one when cost is minimized. And, scale inefficiency measures the difference of the output between the actual level and the one when profit is maximized.
Our data covers the industries of agriculture, mining, manufacturing, construction, wholesale trade, retail trade, accommodation, finance and insurance, communications, and recreational services. The inputs are IT capital, non-IT capital, and labour, both quantities and price deflators. The IT capital includes hardware and software. Thus, we not only can evaluate IT productivity of the entire country, but also that of different industries, as well as the dynamic relationship between IT hardware and software.
Keywords: productivity, software, and productive efficiency
Download FULL PAPER
Back to Contents |
