Policy makers in the public sector are often faced with requests for financial and other support of investment projects and incentive schemes. Frequently, such requests are accompanied by or require economic impact analyses of some sort. Economic impact assessment of investment projects can be undertaken at various levels. At one level, decision makers are interested in the financial viability of the investment project, in other words a comparison of income and expenditure. Taking a broader view, the challenge is to assess the impact of the proposed investment on the economy in which it takes place. Often, rather wild statement can be found in the media in which it is, for example, argued that one job is created for every 12 foreign visitors or so many Rands invested. Although such statements are appealing to the general public, decision makers need to go beyond these aggregate effects and extent the analysis to a more disaggregated level. For example, what will the effects be on the different economic industries? Will these jobs be created for highly skilled or unskilled labor?
The impact of an economic stimulus on specific institutions or industries can usefully be analysed with input-output analysis or social-accounting matrix (SAM) based models. These models use a database or snapshot picture of the economy, and then basically multiply the stimulus with the relevant institution's or industry's output multiplier. However, these analyses rely on strict assumptions, for example production technologies remain constant, which ignores any dynamic effects such as substitution between labor and capital, and a non-substitutability between different types of labor such as skilled and unskilled labor (Holub and Tappeiner 1989).
More specifically, in terms of employment it is often assumed that the average employment-output ratios of the relevant industry apply for all sectors that will indirectly receive a boost as a result of the production activities at hand. If substantial evidence exists of economies of scale and many economic observers have noted the recent phenomenon of jobless growth an alternative specification of the relationship between a change in output and the associated change in employment becomes critical.
Therefore, a time series regression analysis approach will be followed in analyzing the impact of output on labor demand or employment. Apart from generating employment output elasticities, so necessary for a more appropriate application of input-output or first generation SAM based modelling, with this approach it is possible to allow for phenomena such as input substitution and jobless growth, as well as other structural changes to be examined. The study is outlined as follows: the next section summarizes the economic determinants of labor demand. Section three explains the methodology followed in the study, while section four briefly described the econometric techniques used. Section five describes labor demand in South Africa, and section six presents the results of the empirical analysis. Section seven provides some conclusions.