Some Issues in Traffic Forecasting for Privately Financed Infrastructure
KROES E P, Hague Consulting Group, The Netherlands
Privately financed infrastructure has attracted a lot of interest in the last decade. Tolled motorways, bridges, tunnels, urban ring-roads and other infrastructure works are now increasingly familiar everywhere in Europe, and it is likely that much more o
Privately financed infrastructure has attracted a lot of interest in the last decade. Tolled motorways, bridges, tunnels, urban ring-roads and other infrastructure works are now increasingly familiar everywhere in Europe, and it is likely that much more of this will be coming. The type of traffic forecasting required by such privately financed infrastructure is in many respects similar to that used for the more traditional government financed infrastructure, but there are also clear differences.
The typical forecast for government financed infrastructure consists of a best estimate of the "longer term equilibrium" demand level, which serves as an input to a societal cost/benefit analysis or a broader mulficriteria analysis. The outcome of this process is one of the many factors influencing the final political decision whether or not a specific project will be built. The plan horizon here is typically 15 to 25 years ahead in time.
For privately financed infrastructure, the emphasis is much more on predicting the expected revenue stream, with pricing and its predicted effect on demand as key issues. The emphasis is on the expected revenue stream, particularly for the first years of operation, in order to determine the key economic indicators to guide the financial decision process.
There are a number of important issues that need to be addressed when forecasts for privately financed infrastructure are prepared. These include:
* the accuracy of the forecasts: any forecast is likely to be wrong, but the error margin clearly needs to be minimised; how can this be done?
* the uncertainties involved in the forecasts: what are the chances that the revenue will be substantially different from what was predicted due to uncontrollable external factors?
* the time profile of the forecasts, or the build-up of demand: how quickly will the potential users of the new facility actually use it?
Each of these issues have consequences for the methodologies to be used in the forecasting process. In the following sections of this paper we will discuss these issues and their methodological consequences each in turn. This will be done in the light of the current state-of- practice in demand forecasting for major infrastructure links.
Association for European Transport