INFRASTRUCTURE separation is a key plank of the European Commission's strategy to restructure the railways in the member states and make them more market-responsive by opening up the national networks to competition.
When the Commission convinced EU Transport Ministers that the rail sector needed restructuring to keep it from falling into expensive irrelevance, one of the goals was to make rail operations more transparent financially.
To this end, the Commission sought to ensure that budgetary support only went to infrastructure and to support social activities such as suburban or regional passenger services. Government funding should not be used to subsidise national carriers or to suppress competition.
The complex deregulation package has evolved over the past 15 years. Infrastructure separation in particular has proved especially difficult to implement. Indeed, the legal requirements attached to the term 'separation' have changed markedly since the beginning.
The original concept primarily involved clarification of the railway's accounts, so that infrastructure expenditure was clearly distinguished from the cost of train operations. Attached to this was the requirement that other selected operators would be permitted to compete with the national railways, enjoying equal priority for access to the network, and paying a 'non-discriminatory' access charge.
Over the years, the range of markets open to competition has been expanded. Subsequent EU directives mandated that either the rail infrastructure should be managed by an agency separate from any of the train operators or that the access regime (path allocation and charges) should be managed, or at least supervised, by a separate agency.
Discrimination has proven to be a slippery concept, because the political and economic meanings can be in conflict. There is little open dispute that a government should not use the access regime or financial support to discriminate in favour of its national carriers. But price discrimination - in the sense of using an operator's ability to pay as a way of recovering fixed infrastructure costs - is one of the fundamental tools of railway pricing. The challenge has been to achieve an efficient pricing structure without permitting political discrimination.
Degrees of separation
Most governments have chosen to implement institutional separation by setting up an infrastructure manager separate from all operators, which can establish its own access priorities and charges (subject to government oversight) and handle its own timetabling and day-to-day train regulation.
Some governments (notably Germany) chose to separate the infrastructure management from the national train operators, but to keep them all as part of a larger holding company. France and Finland chose to establish separate infrastructure managers, but to have this entity contract some (initially all) operating functions back to the national train operator. In yet another variation, Hungary opted to keep infrastructure and operations under railway control, but to establish a separate agency for setting access priorities and charges.
Most countries have maintained public ownership and operation of their rail infrastructure. Only the UK and Estonia have attempted to privatise their infrastructure. Many countries keep some sort of national or local control of passenger service provision by issuing franchise concessions or fixed-term contracts to both public and private operators.
Freight train operations have been deregulated to a greater degree, with the emergence of several private companies that run across national boundaries using open access rights (p31). In most cases the national railways still handle cross-border freight traffic through co-operation agreements, but Germany's Railion has taken over operators in the Netherlands, Denmark and Italy.
Conditions of access
The key conditions for separating infrastructure are easily understood. Physical access to paths must be fair and politically non-discriminatory, and there must be a set of charges for the use of the track that encourage efficient use while avoiding discrimination among similar users. However, some price discrimination is considered acceptable if it promotes efficiency and does not unduly discourage use of the network.
Of these conditions, physical access has turned out to be relatively easy to define and oversee. The setting of efficient and non-discriminatory access charges has proved more difficult. Recent studies on the issue of access charging have identified a number of questions that have emerged as the access regimes begin to work in practice1.
What has evolved so far is a patchwork of different levels and structures for access charges (Fig 1). Access charges for passenger trains range from about €0·5 per train-km in Sweden to around €4 in France and Germany. Freight access charges range from €0·4 per train-km in Sweden to more than €8 in Slovakia. Costs do vary across Europe, but not by a factor of up to 20, so what is the reason for this variation?
Marginal or full cost?
Probably the most important factor is the infrastructure authority's total income objective. Although the Commission's recommended approach is for access charges to cover marginal costs, and for governments to make up the difference, Directive 2001/14 does allow 'mark-ups' on marginal costs if governments want to shift some or all of the financial burden from the taxpayer to the rail network users.
Fig 2 shows that current cost recovery ratios range from 5% or less in Norway and Sweden to a full 100% in the Baltic states. A rough rule of thumb has been that recovery of marginal costs (including renewals) should be around 20%, so some countries are actually collecting less than their marginal costs while others are attempting to achieve full financial cost recovery.
Clearly the level of access charges will be related to the cost recovery target. Equally, the financial viability of the infrastructure provider will be more dependent on the reliability of government funding where the cost recovery ratio is low. The countries of central and eastern Europe are heavily grouped in the upper range of revenue targets in Fig 2, indicating that their level of access charges will tend to be higher than in other parts of the EU.
A second factor to be considered is the relationship between passenger and freight charges. Are governments trying to 'tax' the unsubsidised freight operators to support subsidised passenger operations, in the hope of reducing the demand on the public budget at the expense of freight shippers? Fig 1 shows that all 10 of the CEE countries listed have freight access charges higher (sometimes significantly higher) than their passenger access charges.
Significantly, of the 13 western European countries, only six have higher freight access charges; in the other seven the passenger charges are higher than those for freight. Moreover, every one of the CEE railways has a ratio of passenger to freight access charges that is lower than in any of the Western countries.
Overall, it is hard to escape the conclusion that most CEE countries are still mired in the former socialist approach of cross-subsidies from freight to passenger operations. In an open market, of course, this approach will inevitably harm the competitive position of rail freight operators in the fight against road transport.
The importance of this goes beyond national policies as the over-riding goal of the EU transport deregulation policy2 is to increase rail's modal share of the freight market. Taken across the enlarged Union, this policy relies on preserving rail's modal share in the new member states whilst growing long-distance international rail services between the old and new member countries as their markets integrate.
Simple or complex charging
The other major difference in the structure of access charges is between simple and two-part pricing regimes. Of the 23 countries listed in Fig 1, 16 employ simple access charge regimes and only seven a more complex two-part structure.
Simple charges are directly related to actual use of the network, usually employing a charge per train-km or per gross tonne-km (or some combination of both). These charging rates can still vary between freight and passenger operators. Charges per train operated or per station stop are sometimes used as well. In some countries, for example Germany and France, the usage charges are also related to particular routes, being higher for high speed lines used by TGV or ICE services, and much lower for low-density secondary lines or rural branches.
The alternative is a two-part regime, with fixed and variable elements. The fixed charge is determined by some relationship of the operator's planned needs to the fixed costs of the network (usually capacity) and the variable element is based on actual measures of use. Most two-part charge regimes base the fixed component on the number of scheduled trains or path-km, with the variable part based on train-km operated or gross tonne-km hauled.
Simple charges are clearly appropriate for collecting marginal costs, most of which vary directly in line with easily-measurable usage statistics. Simple charges are less effective for boosting the proportion of fixed costs to be collected, because they tend to drive down infrastructure use below the most efficient levels. Two-part charge regimes do a better job in this respect, but tend to act as a barrier to competitive entry where the fixed component is relatively high.
Learning the lessons
Although the implementation of EU rules on access charges has resulted in the development of quite complex systems, some broad lessons are beginning to emerge from experience so far.
Critically, access charge regimes not only play a role in covering costs, they also strongly influence the way that the infrastructure is used, especially when the operators are competitive and market-driven. Access charges are actually price signals, and they must be formulated as such. For example, they should reinforce government objectives for rail development in competition with road haulage.
Where the infrastructure is used by several types of operator, a mixed access regime is probably better than a single type of charge. Commuter services, for example, are not usually subject to market competition, except when the concession itself is up for renewal. Thus, as competitive entry is not an issue, a two-part charge can be safely used to provide efficient price signals in these markets. Moreover, these concessions rarely cover more than one country or regulatory jurisdiction, so they can adopt a unique charging regime without undermining competition. Much the same can be said for high speed lines, where the infrastructure is generally used by a single operator.
In contrast, increased competition across international boundaries is an explicit objective for the freight sector, so simple access regimes with limited variation between networks will be critical. An infrastructure manager should therefore consider using two-part charges for suburban and high speed passenger services and simple charges for freight operators. Where competition in the conventional inter-city passenger business is an objective, then these operators should be offered simple charges as well.
Clearly, if Europe's railways are not to become an irrelevance in the continent's wider freight transport market, an effort to develop simple, relatively consistent access charges for freight trains deserves a high priority. Consistency in the structure and level of access charges for suburban and high-speed passenger systems is much less important.
If there is to be any chance of achieving the EU goals for rail market share, cross-financing of passenger services from freight charges needs to end in central and eastern Europe, in order to drive down the level of freight charges. Moreover, the governments in such countries must accept the responsibility of funding fully the public service obligations they impose from their public budgets.
The final, fundamental, lesson is that although there is increasing convergence on the theoretical basis for infrastructure charging, there still exists no base of Europe-wide consistent, reliable and publicly-available data for actually measuring infrastructure costs (however they are defined) and relating them to measures of use.
Until standards for collecting and reporting appropriate data to public authorities can be developed and agreed, Europe's railways and their governments cannot be confident that the full potential value of their rail networks is being realised. They will have an inadequate basis for determining if train operators are being charged a fair price for using the infrastructure and little or no idea of how efficient their infrastructure managers are. So far, this is an area of regulation that has been neglected at the European level.
* Lou Thompson is Principal of Thompson, Galenson & Associates, and Stephen Perkins is Principal Administrator for the European Conference of Ministers of Transport. This article reflects only the views of the authors, and does not necessarily reflect any of the policies and positions of the ECMT and its member states.
1. Railway Reform and Charges for the Use of Infrastructure, OECD, Paris, 2005. (The figures and numbers used in this article are taken from that report.)
2. European Transport Policy for 2010: Time to Decide. European Commission White Paper, COM (2001) 370.