Application of MDCEV to infrastructure planning in regional freight transport

The main objective of the paper is to develop a model capable of evaluating the societal impact of rail infrastructure investment in Argentina, using a Multiple Discrete Extreme Value Model (MDCEV) estimated on Stated and Revealed preference data. The decision modelled is the mode and port choice at a planning level, where multiple alternatives can be chosen simultaneously. The relevant variables were the Free Alongside Ship (FAS) price, freight transport cost, travel time and lead time, including non-observed heterogeneity in the modelling. As a consequence, the willingness to pay measures that are used for the cost benefit analysis become non-deterministic. To include this effect simulated WTP measurements were included and compared to a deterministic and risk based approach. Two projects were tested and both showed that the deterministic approach gives higher Benefit/Cost ratio. This paper raises the concern that if non-observed heterogeneity is not considered in project evaluation it may provide misleading results and potentially lead to wrong investment priorities for the public sector.