EXECUTIVE SUMMARY
1. This research examined the structure, ownership, control, planning, funding, operation and regulation of railways in Ireland, Northern Ireland, Hamburg/Schleswig-Holstein, Denmark, Sweden and New Zealand to identify lessons for Scotland's railways. All are peripheral or isolated railway networks serving populations of fewer than ten million and with substantial sections of single track but one or more urban commuter lines. These networks are managed in different ways: convergence is appearing in some areas but in others a variety of approaches appear to work successfully.
2. Only in Ireland and Northern Ireland are there single controlling authorities which are responsible for both rail and bus transport in both urban and rural areas. The other networks are subject to at least two tiers of cross-border, national, regional, local or urban government influence and control, and the responsibility is divided between government and industry bodies in different ways in each network. In large cities, however, rail timetables, fares and ticketing are increasingly integrated with all other modes, where there is accordingly a need for rail planning and investment to be integrated with other modes.
3. European law expects separation of railway operations and infrastructure, to facilitate open access to competing commercial services, and tendered competitions for non-commercial services justified on social grounds. In practice, almost all rail services face pressure to increase quality and reduce prices through competition from other modes.
4. Some rail freight services can operate commercially, particularly over longer distances, and their provision is left in the hands of operators who may be in the private sector. In contrast, few passenger services can operate commercially and there is little scope for competition between them "in the market". The principal reasons are the levels of fares sustainable in the market or permitted by regulation, lack of infrastructure capacity, desire not to abstract revenue from supported services and the need for fares integration.
5. Most passenger services must therefore be specified and supported by public bodies with competition, if any, through a franchising process, "for the market". With few services commercially viable, access charges have little influence on the behaviour of either operators or infrastructure managers, although a performance regime can ensure that operators do not bear risks related to infrastructure over which they have no control.
6. It is difficult to identify either the effective level of support for individual rail services or even the underlying levels of support and investment in the rail networks. We estimate that, across Scotland and the comparators, the average proportion of passenger train operating costs recovered from fares varies from 40% to 70% and the average level of passenger support varies from 2p/passenger-km to 17p/passenger-km.
7. Patterns of fares regulation vary widely. Some networks have a rigid structure of largely distance-related fares with few discounts, but regulation has been removed from inter-regional fares in Sweden and is largely irrelevant on Anglo-Scottish services, where the operators offer large discounts on the regulated fare on most services. Debate centres on the availability of discounted tickets rather than the level of the regulated fare. In urban/suburban areas, in contrast, there appears to be convergence on zonal fares allowing travel on all modes, although smart cards may in future allow prices to be varied to manage demand.
8. Capacity on single track rural networks is often adequate or easy to increase through use of loops and longer trains, but rising demand in urban areas is creating a need for costly new capacity. This often raises the question of whether it should be added to the existing network or instead through a largely-segregated commuter network, a separate metro system, a light rail network or even by augmented bus services.
9. Responsibility for regulation of the railways lies with different government, regulatory and industry bodies in each network. No consensus either for or against any particular regulatory model has emerged. A key issue, however, is the need for bodies responsible for making planning, procurement and regulatory decisions to have a critical mass of workload and be able to recruit and retain capable, empowered staff.
10. Privatisation of the railway infrastructure has been tried in New Zealand but abandoned after the network gradually deteriorated. Difficulties include specifying and monitoring the condition and capability of the infrastructure, planning publicly-specified services, separating the costs, funding and outputs of underlying operations maintenance and renewal ( OMR) and enhancement projects, and carrying out these projects on a working railway.
11. Privatisation of railway operations through concessioning or franchising has been introduced in all the networks except Ireland and Northern Ireland. Risk transfer arrangements vary widely, from a concession in which the private sector provides infrastructure and trains and takes revenue risk to a management contract with payment linked to performance. Franchises vary widely in size but are normally designed around services using distinct fleets or serving distinct areas. To be effective, however, franchising processes must be designed to avoid a situation in which franchisees retain profitable services but abandon unprofitable ones.
12. The principal barrier to entry to operators wishing to compete for franchises is the need to obtain suitable rolling stock, but leasing markets have emerged or been created in Scotland, Denmark, Sweden and, to a lesser extent, Germany. The need for access to other services in monopoly supply is generally being addressed through effective railway regulation and competition policy.
13. Trading between infrastructure managers, or between operators, and the development of railway supply industries, has allowed a range of subcontracting to procure services not provided in house or to deal with workload peaks. Subcontracting should not normally be mandatory, as it may not be cost-effective, but it presents no particular difficulties provided that workable specification and management arrangements can be put in place.