Dynamic modelling of demand risk in PPP infrastructure projects : “The case of toll roads”

Infrastructure is the main driver of prosperity and economic development. To fill the gap between increasing demand for infrastructure and supply, the role of the private financing has become increasingly critical. Concession contracts in which the investment cost is recovered via payments from the end users are the most dominant among all PPP types. Although this mechanism has been seen as an efficient way to achieve infrastructure projects in terms of realising the project on time and to budget, the demand risk faced in the operation stage has heavily limited this efficiency. Evidence has shown that shortfall in demand can seriously jeopardize the scheme’s viability. Demand is dependent on a range of interrelated, dynamic factors such as economic conditions, willingness to pay and tariff for using the facility. In addition, uncertainty is an inherent aspect of most demand-underlying factors which makes demand estimation subject to high level of uncertainty. However, this uncertainty is largely ignored by modellers and planners and single demand estimate is often used when evaluating the facility. Given the threat to the project success resulting from potential variation between predicted and actual demand, it is believed that a demand risk assessment model is essential. This research is therefore devoted to developing a system dynamics model to assess demand risk by capturing the factors affecting demand and their relationships and simulating their change over time. A system dynamics based conceptual model was developed for mapping factors affecting demand for service provided by a typical PPP concession project. The model has five Causal Loop Diagrams (CLDs) which include: socio-economic, public satisfaction, willingness to pay, competition and level of fee. Based on the developed conceptual model, a quantitative simulation model for assessing traffic demand in toll road projects was developed. This model has six sub-models which are: socio-economic, public satisfaction, willingness to pay, competition, toll and expansion factors sub-models. With the use of case study of M6 toll roads (UK), it was demonstrated the potential application of SD as a tool for the assessment of demand risk in toll roads. Univariate and multivariate sensitivity analysis, as well as risk analysis using Monte Carlo approach, were conducted using the developed SD model. Univariate sensitivity analysis helps identify the significance of the demand underlying factors when they change individually. Toll was identified as the most critical factor affecting toll traffic demand followed by congestion on the alternative un-tolled facility. Multivariate sensitivity analysis showed how demand changes when several factors change. Four scenarios were developed to show the impact of change in conditions and policies on the level of traffic. Monte Carlo simulation, on the other hand, provided level of demand with a range of confidence intervals. Providing such estimates of the expected value and the confidence level offers useful information throughout their ranges and creates overall risk profiles by providing the probability of achieving a specific result. The main contribution of the research is in the development of a system dynamics model as a tool for assessing demand in PPP projects and informing decision making, which is new to the area of demand risk modelling