Estimates of marginal infrastructure costs for different modes of transport

One component of optimal prices for infrastructure use is the marginal cost of maintaining and operating infrastructure. While extensive studies on optimal congestion and environmental charges as well as the respective cost estimates are available much less attention has been paid to the estimation of marginal infrastructure costs probably due to the assumed lower quantitative importance for pricing compared to environmental and congestion costs. This paper presents results from a set of studies on marginal infrastructure costs for different modes of transport. It is based on research previously undertaken for the European Commission within the UNITE project (Unification of Marginal Costs and Accounts for Transport Efficiency in Europe). The studies employed different methodologies for estimating marginal costs ranging from econometric approaches up to engineering based methods. The focus of the analysis is on road and rail, however, the paper includes also results for an airport and for seaports. The main finding from the methodological point of view is that the "one" ideal methodological approach to estimate marginal infrastructure costs does not exist. Econometric approaches are based on observed behaviour of costs and cost drivers. It is obvious that the actual or observed costs do not always follow technical needs resulting from the use of infrastructure, i.e. do not necessarily reflect true marginal costs. In comparison, marginal costs derived with engineering-based methods are built on measured technical relationships, but which are not necessarily reflected in actual spending. Both econometric and engineering based approaches require a considerable amount of high-quality data with a demanding level of detail so that under this aspect one cannot express a preference for one of the two. The experience with econometric approaches was that in absence of data on axle-load km (for road and rail) the inclusion of transport performance indicators for different vehicle categories or types of transport as explanatory variables cause serious multicollinearity problems which can only be solved by constructing aggregate indicators with the consequence of restrictions to interpretation of results. Engineering based approaches which use axle-load km face vice versa the problem that marginal costs per axle-load km have to be translated back into marginal costs per vehicle type. The research presented in the paper provides evidence that for rail tracks and road infrastructure it is mainly the cost of maintenance, repair and renewal that vary with traffic volume. For terminal infrastructure such as ports and airports it is staff costs which varies in the short run with traffic. For rail tracks and road infrastructure the main cost drivers identified are traffic load, especially measured by weight indicators such as gross-tonne km and axle-load km, infrastructure characteristics such as number of bridges, tunnels, electrification etc., age of infrastructure and maintenance history. For terminal infrastructure where staff costs form the major category of marginal infrastructure costs the traffic load (measured as throughput in ports and as aircraft movements and departing/arriving passengers at airports) is again the main cost driver. In addition, the studies provided evidence that the season, the weekday and the salaries' arrangement have to be considered for analysing operation costs of terminal infrastructure. Both the econometric and the engineering based approaches mostly provided results which are consistent with the u-shaped marginal cost curve suggested by neoclassical economic theory. However, in many cases the detected non-linearities were rather weak in the relevant range of traffic variables. No uniform result was obtained with respect to the question which branch of the "u" describes marginal infrastructure cost behaviour. The quantitative results of the studies are widespread and indicate the need for further research in the field. This holds also true for airport infrastructure and waterborne infrastructure were so far only few studies are available.