Transport modelling – fact, forecast or fiction?

‘Transport modelling – fact, forecast or fiction?’ was the topic of a well-attended meeting of the Transport Planning Society at which I was a panelist. I argued that there was occurring quite a lot of change in travel behaviour as we moved into the twenty-first century – not least the Peak Car phenomenon – which made the task of the modeller more difficult. Modelling of any kind assumes continuity between past and future, that past relationships (estimated as elasticities) will apply in the future, subject to changes in parameters exogenous to the model, such as growth in GDP, population and oil prices. If behaviour is changing, the best approach is to widen the range of forecasts by adopting scenarios which allow the model to explore the impact of a wider range of travel behaviour. An example is the generation by Department for Transport modellers of road traffic forecasts based on five scenarios applied to the National Transport Model.

I also drew attention to the experience of the Actuarial profession, which after the failure of a life assurance company had prompted a government inquiry,  had put in place formal standards for actuarial analysis and a means for professional oversight of compliance. One standard deals with modelling, the language of which is quite general and would be relevant to other kinds of modelling, including transport modelling. So the actuaries’ arrangements show that it would be possible to put in place formal standards for transport modelling. However, for this to happen, there would probably need to be some kind of scandal, as happened to the actuaries.

One kind of scandal involving transport modelling has occurred in Australia, where a number of privately-funded toll roads have experienced usage far below the forecasts made when investors were approached to finance construction. This has resulted in litigation that in at least one case resulted in the transport consultant responsible for traffic forecasts paying out $200m. Were something similar to happen in Britain, I would expect a call to put in place standards for transport modelling.

My fellow panelists had their own concerns and solutions to achieve better transport modelling. My feeling from the meeting as a whole is that there is a needed to  review systematically the current state of the art and to identify ways to improve. I hope the Transport Planning Society might act as a thought-leader, given the centrality of modelling to planning.