require ad hoc judgements about the current cycle in order to keep the results within reasonable
bounds.
When applying the first method, incorporating a stochastic production function-framework, a
distinction needs to be made between “potential” and “normal” output. Modelling potential
output as opposed to normal output requires the estimation of potential levels of factor inputs.
“Normal” output, defined as the production level with the current quantities of factor inputs and
operating at a “normal” or trend rate of utilisation, is usually obtained by smoothing the various
components of the production function (Turner et al. 1996).
It should be noted that from the point of view of macroeconomic analysis, the structural
production-function approach is preferred. The most important limitation of any smoothing
method is that it is largely mechanistic and ignores all structural properties associated with
production. Aspects such as the availability and quality of factors of production, their
productivity, the production technology and technical progress, and all other exogenous
influences are not taken into account. The trend output growth projected by time-series
methods may be inconsistent (too high or too low) with what is known or being assumed about
the growth in capital, labour supply or factor productivity. The trend growth may also be
unsustainable because of the ignored inflationary pressures.
A structural production function approach therefore has the advantages of overcoming the
above-mentioned shortcomings, incorporating the role of demand pressure on employment and
inflation and allowing for consistent judgement on some of the key elements. The production
function approach explicitly models a production technology in terms of factor inputs, factor
technology and to some extent the role of technical progress. Potential output is then
determined as the level of output that results when the factors of production and total factor