Transport activities give rise to environmental impacts and accidents. In contrast to the benefits, the costs of these effects of transport are generally not borne by the transport users. The internalisation of external costs means making such effects part of the decision-making process of transport users. This handbook outlines a model for the internalisation of external costs which will serve as a basis for future calculations of Infrastructure charges.
Update of the Handbook on external costs of transport (January 2014)
This new 2014 Handbook on external costs of transport continues to present the state of the art and best practice on external cost estimation. In comparison to the 2008 Handbook, it takes into account new developments and progress in the following fields:
- Large new databases on noise, accidents and emission factors,
- New and updated models,
- Updated estimates of important input parameters,
- Research identifying additional health effects
- Case studies and marginal cost calculations.
The 2014 Handbook also integrates infrastructure costs – previously tackled in a separate report –and provides updated and more detailed country and area specific estimates of marginal external cost estimates.”
Updated unit values for congestion costs – rail
The study of the existing literature did not reveal many new sources (as compared to the 2008 Handbook) of marginal congestion or scarcity cost estimates for rail, air, or water transport that could be recommended as a best practice methodology. However, it is obvious that some national methodologies for pricing infrastructure access do take account of the variation of traffic flow e.g. according to time of the day and type of path (e.g. for rail), which suggests that the scarcity of slots at peak hours has an impact on the level of charges.
For rail transport, recent overviews of national practices in charging for infrastructure access have been carried out by the International Transport Forum (2008) and in the DICE Database (2012). Annex G includes an overview of access charges presented in these two sources. The introduction of the ERMTS (standardised signalling developed to be used within Europe, but used elsewhere as well) has had a major impact on the reduction of delays in rail transport, both freight and passenger. The minimum headway between trains on some heavily used lines could be reduced to 2-3 minutes using the ERMTS level 2 (UNIFE, 2012). If true marginal congestion costs for rail transport
were to be calculated, these facts must also be taken into account.
The marginal cost estimate for freight rail congestion as contained in the most recent version of the Marco Polo calculator (Brons and Christidis, 2013) is €0.2 per 1000 tkm (average for EU27, in 2011 prices). This number is derived from the studies reviewed in the 2008 Handbook. The average is calculated by assuming equal freight rail congestion costs in most EU countries at the level of €0.1 per 1000 tkm. For Italy, the estimated unit cost is €0.25, for Germany and France €0.4, and for Belgium and the Netherlands €0.5.
Jansson and Lang (2013) have developed a new methodology to evaluate the external delay costs in rail transport. In the application for passenger transport in Sweden, the authors estimate, how the marginal cost-based charges (initially limited to external costs for wear and tear, maintenance, emissions etc.) would change if delays due to additional departures were also taken into account.
For example, if an additional departure of a commuter train leads to a delay of two minutes in the network shared with high speed trains, the authors estimate the marginal external cost effect of this delay to correspond to a 25% increase in the commuter train fare for this additional journey, and a 5% increase in the fares for high speed trains. Overall, Jansson and Lang (2013) suggest that charging for delay costs should be introduced for the operators in the market that cause large negative external effects and whose customers have low valuation of wait and delay time (operators of commuter trains, in the example above). However, introducing such pricing schemes in practice may be difficult.
Updated unit values for air pollution costs
Marginal infrastructure costs
The topic of rail infrastructure costs was not included in the 2008 Handbook. The calculation of these costs however has important policy implications. In the course of railway liberalisation in Europe, the network operators were obliged to reveal information about the costs that form the basis for the determination of network access charges (Directive 2001/14/EC). These charges must be based on a transparent methodology, with the direct cost of operating the railway service (plus a reasonable rate of return) forming a lower bound for such a charge. The correct differentiation of the charge for different types of users is only possible if the marginal
costs are calculated, that account for the specific contribution of different users to the total costs of infrastructure wear and tear. Most recent joint efforts in order to establish methodological recommendations for the Member States in this respect were undertaken in the course of the CATRIN project (Wheat et al., 2009). The starting point for the top-down calculations is the following representation of the marginal cost:
Marginal cost = (Average cost) x (Cost elasticity)
First, the relevant cost must be identified. Most studies concentrate on the maintenance costs only.
permanent way costs,
signalling and telecoms costs,
electrification and plant costs.
More rarely, renewal costs are also considered. Network-wide overheads, however, are not relevant for determining the optimal infrastructure use charges. The cost elasticity can consist of several components, depending on the data availability and policy needs. The components of elasticity could quantify the impact on the total cost of:
total amount of traffic (track usage)
type of track (electrified or not; dedicated freight or mixed line)
type of train (passenger, freight; regional, intercity, etc.)
In CATRIN case studies for Great Britain, Austria, Sweden, Switzerland and France, the cost estimates are differentiated between passenger and freight and the cost elasticities are given only for three traffic density ranges (in tonne-km per annum). Cost elasticities are generally in the range of 0.1-0.35, meaning that marginal cost-based prices will require substantial mark-ups if the full cost of maintenance and renewals is to be covered, let alone a contribution made to investment costs (Sanches-Borras et al. (2010)).
Overall, the following findings from the literature are important to note before some results are presented:
Impact of traffic density. Many studies refer to the U-shaped form of the traffic elasticity, meaning that the reported econometric estimates of this elasticity decrease with density at low density values, and then increase when density reaches some threshold value (often close to the mean). According to some recent findings (Gaudry and Quinet (2013)), this effect is not always present. What remains true is that the background traffic amount is a very important factor for the level of marginal costs.
Ratio of passenger to freight marginal costs. Most studies find that the marginal costs for freight trains are substantially lower than for passenger trains (1.5 to 7.5 times, according to the estimates in Wheat et al. (2009)). Gaudry and Quinet (2013) name the following reasons for this phenomenon: repairs on passenger-only lines must happen much faster and are thus more expensive; segments with a large proportion of freight trains do not require a high level of quality; due to shortage of funds, freight lines are more likely to get cheaper preventative maintenance rather than more expensive curative maintenance.
Type of econometric model. Modern econometric techniques allow the use of estimating models with nonlinear parameters. This may lead to a revision of older estimates using exclusively linear models.
Power function for load damage. For road transport, the fourth-power law (see next section for an explanation) is applied to allocate damage costs to vehicles with different axle load. In rail transport, the dominant view is that the relation is simply linear. Gaudry and Quinet (2013) present some indication that non-linear relationships with the power factor greater than unity may be plausible, but there is no strict proof of this so far.
Link to 2014 Handbook on external costs of transport http://ec.europa.eu/transport/themes/sustainable/studies/doc/2014-handbook-external-costs-transport.pdf