The correlation of externalities in marginal cost pricing: lessons learned from a real-world case study

Negative externalities cause inefficiencies in the allocation of capacities and resources in a transport system. Marginal social cost pricing allows to correct for these inefficiencies in a simulation environment and to derive real-world policy recommendations. In this context, it has been shown for analytical models considering more than one externality, that the correlation between the externalities needs to be taken into account. Typically, in order to avoid overpricing, this is performed by introducing correction factors which capture the correlation effect. However, the correlation structure between, say, emission and congestion externalities changes for every congested facility over time of day. This makes it close to impossible to calculate the factors analytically for large-scale systems. Hence, this paper presents a simulation-based approach to calculate and internalize the correct dynamic price levels for both externalities simultaneously. For a real-world case study, it is shown that the iterative calculation of prices based on cost estimates from the literature allows to identify the amplitude of the correlation between the two externalities under consideration: for the urban travelers of the case study, emission toll levels-without pricing congestion-turn out to be 4.0% too high in peak hours and 2.8% too high in off-peak hours. In contrary, congestion toll levels-without pricing emissions-are overestimated by 3.0% in peak hours and by 7.2% in off-peak hours. With a joint pricing policy of both externalities, the paper shows that the approach is capable to determine the amplitude of the necessary correction factors for large-scale systems. It also provides the corrected average toll levels per vehicle kilometer for peak and off-peak hours for the case study under consideration: again, for urban travelers, the correct price level for emission and congestion externalities amounts approximately to 38 in peak hours and to 30 in off-peak hours. These toll levels can be used to derive real-world pricing schemes. Finally, the economic assessment indicators for the joint pricing policy provided in the paper allow to compare other policies to this benchmark state of the transport system.