Differentiated road tolls: evidence from modelling studies in Italy and the UK

Road tolls are increasingly applied in several European countries, including those where the use of the road network had been traditionally free of charge. Different tolls are often applied to different categories of user although the rationale behind the differentiation is not always clearly established. Transport economists advocate the use of pricing with the aim of improving efficiency and internalising external costs. From this point of view, differentiation can be used to better adapt tolls to the level of marginal costs. The European project DIFFERENT funded by the European Commission DG TREN has investigated the role of differentiated prices from a theoretical and empirical perspective. Within DIFFERENT, several modelling tests have been carried out to assess the impact of various differentiation schemes. This paper reports the main findings of simulations carried out using network transport models of the Brenner corridor, of the Padana region in Italy and of a number of modelling studies in the UK. The Brenner model and the Padana region models are implemented using the Meplan software package. The Brenner model simulates both modal split and route choice of both passenger and freight demand, while the Padana region model is a pure assignment model. Two parallel segmentations of demand are used in both models, one concerning vehicles and one concerning individuals or transported goods. Several types of freight road vehicles and passenger cars are modelled according to three main dimensions: size, fuel type and standard EURO. In the Brenner model also a segmentation of passenger and freight demand is used according to trip purpose, commodity group and average trip length (Crossing traffic; Short and Long distance traffic). The demand segments are used only to define different behavioural parameters (e.g. value of travel time) and not tolls differentiation schemes. A number of scenarios were simulated with both models, where tolls were differentiated according to various elements: EURO category, freight vehicle size, passenger vehicle occupancy, road type. The simulations suggest that there may be a trade-off between different objectives of charging policies. Improving travel time on the network can entail a reduction of the revenues for the motorway operator, or applying motorway tolls proportional to the level of emissions can give rise to undesired environmental effects. Under an economic perspective, an evaluation of scenarios has been carried out assessing the shadow prices of travel time and pollutant emissions as well. In this way different outcomes could be added to each other. A noticeable result is that social benefit of the same tolling scheme are different in the Brenner case, which is a corridor with some spare capacity and the share of crossing traffic is high, and in the Padana region case, where a complex and congested network is analysed. The paper reviews also previously published results from modelling studies in the UK which have addressed the issue of differentiated road user charges. The most notable of these is the UK Department for Transport’s National Road Pricing Feasibility Study which suggested that a scheme utilising up to 75 different charge levels might perform particularly well. Other previously published studies reviewed here include several of the Multi-modal studies which tested the effect of different charging regimes. The paper also includes results from a modelling study conducted at Leeds University which used a multi-class elastic assignment model of a medium-sized metropolitan area to explore the performance of different types of charge structure (with different degrees of differentiation). The model suggested that highly differentiated charges designed to minimise marginal social costs worked well in terms of their impact on externalities and their generation of revenues. However, charges of this type resulted in a significant amount of trip suppression and, given the effect that this has on consumer surplus, their overall performance was actually inferior to that of simpler schemes such as a fixed charge per unit distance. The implications of this important result are discussed. Conclusions drawn from the comparison of studies in Italy and UK end the paper.