The use of state-of-the-art transport models by policymakers – beauty in simplicity?

Abstract

Transport demand models have a long history of being a major tool in transport policy making. However, whether they are truly used in decision-making processes, and if so, whether the knowledge they provide is actually understood, is questionable. The potential contribution they can make and the importance of such models is not disputed; however, evidence shows that many issues arise with their actual use that severely limits their potential contribution. Based on case study methodology and analysis of the use of models in the transport policy processes in two countries, the UK and Israel, this paper aims to provide empirical evidence of the issues contributing to limiting the potential contribution of models and to make recommendations for better utilisation of the knowledge they can produce. The main conclusion reached is that transport models must be made simpler if they are to contribute more than they currently do to decision-making in transport policy and planning.

Keywords:

• Transport planning
• demand modelling
• model development
• knowledge utilisation

Acknowledgements

The authors would like to thank Omer Zur and Robert Ishaq from the Technion, Israel Institute of Technology for their help in collecting and analysing the data. We would like to thank the British Council’s Britain–Israel Research and Academic Exchange Partnership (BIRAX) programme for making this research possible by funding it. We are grateful to the policymakers, academics and planners who agreed to be interviewed, and to those taking part in two project workshops. We would also like to thank the Technion Institute of Technology, Israel and St Anne’s College, University of Oxford, for hosting the workshops.

Notes

1. We use "policymaking" when referring to the policy process and "decision-making" when referring to its outcome.

2. According to Lee (1973) the "seven sins of large scale models" are: “hypercomprehensiveness” – in simple words, including too much detail or variables; “grossness” – too wide in scope; Hungriness – for data; “wrongheadedness” – explicit or implicit commitment to the status-quo in building and structuring the model (equations); “complicatedness” – which requires the model to be "massaged" (constrained) in order to provide “reasonable” results; “mechanicalness” – the risk of focusing on “solving” the problem on the computer instead of thinking the problem through; and finally Expensiveness – of the cost of modelling efforts.