Reducing or Managing the Forecasting Risk in Privately-financed Projects
M Bright, A Boyce, FaberMaunsell, UK
There is inevitably uncertainty surrounding forecasting demand for new infrastructure, particularly in ?pay as you go? systems, and this uncertainty has long been recognized. This issue has been thrown into stark relief recently by the release of a document by Standard and Poors (?S & P?), analysing forecast and outturn traffic. This report has cast doubt on both the competence and integrity of the transportation planning profession.
The Report concludes that outturn traffic is nearly always lower than that forecast and often considerably so. This can have serious implications for those organisations financing the scheme. S and P suggested a number of ?stress tests? for scheme forecasts which depended on perceived reliability of certain aspects of the project, such as type of scheme (toll or shadow toll) or to the country, rather than the quality of the data or the analytical methods used, or indeed the evaluation criteria for the tenders. The S & P approach could be described as rough and ready. It may result in some worthy schemes being ignored while unworthy schemes are constructed. However, the study did emphasis the difficulties of forecasting traffic and the benefits of having traffic advisors appointed by and reporting to the banking group.
Our experience suggests that a proper risk (probability) assessment is a better way of producing forecasts for use in banking cases. This will produce a central forecast which is equivalent to that produced for concession companies. It will also produce a whole series of forecasts with their probability of being exceeded. The extent to which these forecasts differ from the central forecast depends on the quality of the collected data and the potential variability in forecasting assumptions. These are often scheme specific and offer a superior analytical approach than the ?stress test? values proposed by Standard and Poor.
There is also the issue of so-called ?downside? and ?upside? cases. While the concept behind such scenarios has some merit, their definition is essentially subjective, and the probabilities attached to them arbitrary.
A risk assessment takes account of the potential sources of error in traffic forecasts and combines these probabilities in such a way to determine the overall distribution of the forecasts. These uncertainties may be conveniently divided into Base Year issues and Forecast Year issues.
We will offer a critique of the S and P report and the actual experience of forecasting for privately-financed projects. We will then go on to describe the outcomes of an analytical Risk Assessment compared with scenario definition, and recommendations for the professional conduct of forecasting and dealing with uncertainty.
Association for European Transport