Japan Railway & Transport Review No. 29 (pp.19–23)

Feature: Railway Management and the Role of Government
Separation of Railway Infrastructure and Operations
Vassilios A. Profillidis


The globalization of economic activity and gradual liberalization of the transport market in the 1990s led to major changes in the organizational structure and monopolistic nature of some railways. Separation of railway infrastructure and operations laid the foundations for introduction of competition to railways as well as for their economic rationalization. This article analyses the basic principles of the European transport policy and the objectives of European legislation on the modernization and competitiveness of railways within the transport system. It also examines how railway infrastructure and operations were separated in various European railway networks as well as the structural consequences of the separation. Finally, the railway infrastructure pricing principles as well as the impact of the separation of infrastructure and operations on railway finances and transport demand are discussed.
European Railway Legislation Modernization and Competition
The old European railways urgently needed reforms to offer customers efficient, high-quality, market-oriented services at lower cost. They could not ignore the globalization of economic activity and liberalization of transport markets without remaining hamstrung by out-of-date organizational structure and monopolistic business tendencies.
The finances of traditional railway businesses are inherently non-transparent because they run train operations on their own infrastructure with very high sunk costs. In addition, ownership of infrastructure eliminates any incentive to promote free competition because monopolistic tendencies do not favour entry of other railway operators on the same infrastructure. Unlike other transport modes, railway infrastructure costs in Europe currently account for some 30% of the total operational costs. This was a major factor in the EU drive to separate the accounting of infrastructure and operations because it enables fair comparison with the infrastructure costs of other transport modes, such as roads and airports, which are largely borne by the state.
Against this background and to reverse the declining fortunes of European railways (Figs. 1 and 2), the EU formulated a Transport Policy with five basic aims:
Reduction of various (mainly trans-national) barriers to create single European transport infrastructure
Free market entry and unrestricted operations by transport businesses in markets of Member States
Reduction of environmental impact of transport
Phased abolition of all state intervention in and subsidy to railways and promotion of competition
Standardization of European transport rules, removal of differences in levies and taxes, and elimination of technical barriers
Figure 1: Share of Passenger Transport Market for Each Transport Mode for ECMT Countries
Figure 2: Share of Freight Transport Market for Each Transport Mode for ECMT Countries

Separation of Infrastructure and Operations

EU Directives 91/440/EEC, 95/18/EC, 95/19/EC specify the necessity for separate accounting of infrastructure and operations as the minimum reform; the directives do not actually specify splitting infrastructure and operations into two or more separate business entities. Consequently, Member States have adopted two basic methods to achieve the required minimum: Institutional separation in the UK (Railtrack/Train Operating Companies (TOCs)), and France (French National Railways (SNCF)/Reseau Ferre de France (RFF)), and organizational separation as in Germany (Deutsche Bahn AG (DB AG)/DB Netz).

Institutional separation
This method separates the infrastructure owner and railway operators into autonomous entities with separate capitalization, balance sheets and staff.
The infrastructure owner can be publicly owned as in Portugal (Portuguese Rail Infrastructure Authority (REFER)) and Sweden (Banverket (BV)) or privately owned as in the UK (Railtrack). Even when the infrastructure owner is publicly owned, it must still operate according to the normal rules and laws regulating private businesses. A government-appointed regulator mediates disputes between the infrastructure owner and railway operators.
Railtrack in the UK is the only example of a genuinely privatized infrastructure owner. However, a series of fatal accidents and serious infrastructure problems in the UK coupled with a 1996 DB AG study (Die Deutsche Bahn AG offnet den Fahrweg fur Dritte) suggest that private ownership of railway infrastructure is not necessarily a good idea.
France falls into this category in the sense that the infrastructure manager (RFF) and operator (SNCF) are completely separate legal entities with separate staff, but the relationship is closer than in the UK because SNCF actually operates and maintains the infrastructure based on contracts awarded from RFF.
In Finland, the infrastructure manager is the Finnish Rail Administration (RHK), a department of the Ministry of Transport and Communications.
So far, it seems that the French and Finnish models have avoided the problems of the UK model.

Organizational separation
This method creates separate business units with a large degree of operational freedom. There are two basic patterns:
Business units operating as part of railway operator
This method is used by Belgian National Railways (SNCB/NMBS) and Italian Railways (FS). The units have an independent management and a separate balance sheet but no legal autonomy.
Autonomous business units organized within framework of holding firm
This method is used in Germany. The business units developed into autonomous companies (DB Reise & Touristik, DB Regio, DB Cargo, DB Netz, DB Station & Service) under the holding company DB AG.

Structural consequences of separation
Figure 3 shows the interactions between the subsystems of the reorganized railways and Figure 4 shows the new roles and challenges. Tables 1 and 2 summarize the many questions concerning the new organizational and the factors affecting them.

Figure 3: New Organization Structure of Railways and Interactions between Various Subsystems
Figure 4: New Roles and Challenges of New Organizational Structure of Railways
Table 1: Questions on New Strategic Organization of Railways
Table 2: Factors Affecting New Organizational Model of Railways

Basic Principles of Infrastructure Pricing

Like other transport operators, railway operators must observe the normal rules of business competition, which preclude state subsidy except where the railway operator is obliged to offer unprofitable services, known as public service obligations (PSOs), for social reasons. Excluding PSOs, when infrastructure and operations are completely separated, the railway operators pay fees for using the infrastructure to the owner based on the pricing principles outlined below.
The same pricing principles should apply to all major transport modes throughout the EU. However, structures and pricing levels may vary with transport type and locality due to economic and social differences.
Infrastructure pricing should be based on the principle that the user pays.
Pricing should reflect the infrastructure usage level and method based on full-cost accounting bearing in mind environmental and social impacts, such as the costs of accidents, pollution, traffic congestion, etc.
To prevent competitive distortion, pricing differences should only be allowed when there are actual differences in the cost level and service quality.

Costs of transport infrastructure
Operation of railway infrastructure involves various costs, some of which should be borne by the users. Economic theory divides these costs into fixed costs and direct costs.
Fixed costs includes construction costs, various maintenance costs (lighting, staff) and some other labour costs when there are legal or contractual obligations. In the case of Railtrack, fixed costs account for some 90% of total costs, and about 75% for SNCF.
Direct costs depend on the degree to which the infrastructure is used; if there are no users, this cost is a debit. Most maintenance costs are direct costs. Marginal costs are the additional costs incurred by operating an additional train, etc.
The transport sector has internal costs (costs to users) and external costs (costs to non-users). The latter are costs due to traffic congestion, pollution, noise, safety, etc. Clearly, users of the road infrastructure pay only internal costs but not the external ones, leading to some advantageous disparities in comparison with railways.

Examples of infrastructure pricing principles
By separating the accounting of infrastructure and operations, the EU legislation aimed to make the costs of railway operations more transparent, providing increased management flexibility and gradual introduction of competition. The method adopted for pricing railway infrastructure (RI) was critical to achieving healthy competition.
EU Directive 91/440/EEC stipulates the following obligations:
Only designated railway operators are obliged to pay RI usage charges.
The debit method for RI charges is an obligation of the state.
The charges must not discriminate between RI users.
Although not obligatory, RI charges should be calculated based on parameters such as journey distance, transport type, speed, load per axle, time, etc. The parameters are at the discretion of each Member State.
Railway operators operating exclusively urban, suburban and local services do not fall under the RI pricing obligations. The same applies to Eurotunnel because cross-Channel traffic is interurban and not suburban.

Additionally, EU Directive 95/19/EC stipulates the following obligations:
The RI usage must be balanced within 3 to 5 years after the start of pricing. Consequently, infrastructure expenditure should be balanced in relation to usage revenues after 2002-03. However, this is a difficult goal that only Railtrack and DB AG have achieved, possibly because their access charges are very high.
RI pricing must be non-discriminatory for RI users and this is an obligation of the state.
RI charges shall be calculated according to the principles set out in EU Directive 91/440/EEC excluding PSOs. In the case of PSOs, the state may impose a lump sum usage charge on the operator providing the PSO.
The results of promoting competition, improving services and increasing transport levels shall be evaluated after some period of years for adjustment purposes.

The various tariffication models can be classified into three basic types:
German model with two components:
Constant component (Nenzcard) related to speed, traffic (passenger, freight, express, etc.) and track demand
Variable component (per train km) expressing speed, track demand (0.85 to 1.15), and schedule accuracy (1.0 to 1.4)

French model with three components:
Access component (very disadvantageous to new market entrants)
Reservation component (per train km) to be paid even when track path not used
Circulation component based on factors such as distance, departure time, track condition, etc.

British model with two components:
Fixed charges allocated irrespective of traffic level
Marginal costs (<10%) allocated in relation to actual traffic

In the British approach, the infrastructure owner (Railtrack) always collects at least 90% of the total costs—a financial utopia, but very sensible for a private company.

Percentages of Railway Infrastructure Expenses in Europe

Percentages near 100%
In theory, Germany and the UK both use pricing models where RI revenues and expenses are balanced. However, In Germany, the truth is that various forms of state aid and subsidies comprise about 35% of RI expenses with revenues covering about 65%. For example, some staff costs are subsidized by the federal government. In the UK, revenues should cover 106% (100% expenses + 6% profit) of expenses, but since such a model would quickly wipe out many TOCs, the government provides subsidies, etc., to the TOCs thereby adopting a dual policy of balanced revenues and expenses, and solvency.

Percentages far below 100%
Most EU countries fall into this group as follows:
• Switzerland: 70%
• Italy: 40%
• France: 30%
• Sweden: 15%
• Belgium: 15%

Percentages near 0%
To date, only the Netherlands has applied a zero-pricing policy. However, the General Directorate for Transport of the European Commission with responsibility for competition has enforced a change of this policy.
Sweden is also preparing a modification of its existing pricing in view of the near-zero charges in the freight sector and based on marginal social costs for passenger transport.

Financial Impact of Separation

Except for the Netherlands, the costs of RI usage represent a significant amount of money for railway operators In the UK, the RI usage charges to TOCs were calculated so that the resulting revenues would cover RI expenses. This significant burden on TOCs is tempered temporarily by state subsidy that will drop to zero at the end of the concession period.
In Germany, RI usage charges are also effected in order to cover most of the total RI expenses, thus creating a serious problem for the operating expenses of the railway operator, as well as for the regions of the country, where some competence of questionable effectiveness for the subsidy of regional character railway lines has been conceded.
By contrast, in France, RI pricing is based on the solvency of the railway operator, creating serious deficits for the infrastructure manager.
External Costs and Consequences for RI Pricing
Increasing environmental awareness and a clearer understanding of the environmental impact of the various transport modes are driving the concept that the polluter pays and are resulting in internalization of external costs.
A relatively recent study of 17 countries (15 EU Member States, Norway and Switzerland) showed that the external cost of transport in 1991 was €272 billion (€1=US$0.91) or about 4.6% of the GDP of these countries. This breathtaking amount of money is the social cost of accidents, noise, air pollution, climate change, etc., caused by transport.
Figure 5 shows a comparative analysis of the various transport modes—road traffic accounts for 92.2%, air transport for 5.9%, railways for 1.7%, and inland waterways for 0.3%.
Of course, introduction of common rules for calculating external costs is difficult, but the problem of pricing will be even greater. Internalization of external costs will surely benefit railways while making road transport much more expensive. EU Directive 93/89/EEC aimed to tackle this problem but did not achieve spectacular results. Internalizing the external costs of transport is strongly resisted by special interest groups such as the auto and oil industries and there are no guarantees that it will ever be achieved.

Figure 5: Comparative External Costs of Various Transport Modes

Further Reading
D. Hensher and W. Waters II, Evolution and Revolution—The Changing of Regulation of the World's Railway, JRTR, No. 22, December 1999, pp. 4–11, 1999.
INFRAS, IWW, Effets Externes du Transport (Transporting external baggage), Karlsruhe, 1994.
V. A. Profillidis, Theoretical and Practical Aspects Concerning Land Access to Sea Ports, European Conference of Ministers of Transport, Round Table 113, 1998.
V. A. Profillidis, Railway Engineering, 2nd Edition, Ashgate, 2000.
Energy Policies and Trends in Europe, Biannual Publication, Commission of the European Communities, 1993.
Environment and Transport Infrastructure, European Conference of Ministers of Transport, 1989.
Internalizing the Social Cost of Transport, Organization for Economic Cooperation and Development, 1994.
Railway Infrastructures Charges and Capacity Allocation, European Commission, 1998.
Trends in the Transport Sector, European Conference of Ministers of Transport, 1999.
White Paper—Strategy and Restructuring of the Railways, Commission of the European Communities, 1996.

Vassilios A. Profillidis
Dr Profillidis is Associate Professor of civil engineering at Democritus University of Thrace in Greece. He received his Ph.D. from the Ecole Nationale des Ponts et Chaussees of Paris. He worked as a research associate in the Research Department of the International Union of Railways (UIC) and SNCF. He has published 70 papers in scientific journals and conference proceedings and is the author of Railway Engineering.