Do cities deserve more railway stations? The choice of a departure railway station in a multiple-station region

Highlights

• The literature on the number of railway stations to provide within a particular urban area is scarce.

• Access journey to the station is important in determining the overall (dis)utility from travelling by rail.

• It would not be beneficial to reduce the number of rail stations in the Amsterdam area (currently 11).

• Intercity trains should stop at more stations in the Amsterdam region compared with the current time tables.

• There is a need for an integrated analysis and planning of rail and local public transport services.

Abstract

Promoting the use of rail is an important element in sustainable transport policy. One of the most important decisions to make in planning the railway network is on the number of stations to provide. Stations are the access points to rail services and while each additional station increases rail’s accessibility it also slows the service for those not using a particular station thereby making rail less attractive. In most large cities there are several, often many, railway stations, and understanding how passengers choose a departure station is important for planning the number of railway stations in a particular urban region and for understanding how the urban transport network provides access to different stations. Using the Dutch Railways (NS) customer satisfaction survey, a discrete choice analysis is carried out of the choice of a departure station, out of 11 available, by passengers living in the Amsterdam region. The results confirm the importance of the access journey to the station in determining the overall (dis)utility from travelling by rail. The welfare effect of closing a station in the Amsterdam area is computed by using the logsum approach, leading to the conclusion that it would not be beneficial to reduce the number of stations in Amsterdam. This suggests that increasing the number of stations where trains stop may well be welfare improving for the Amsterdam region. The paper concludes by advocating for an integrated analysis of rail and local public transport services, and for an integrated planning of inter-city rail and urban public transport networks, with focus on the number of rail stations within the urban area.

Introduction

In 2011, the European Commission announced its official transport policy in the White Paper “Roadmap to a Single European Transport Area – Towards a competitive and resource efficient transport system” (EC, 2011) in which rail is to play an important part. One of the targets set by the white paper is: “By 2050 complete a European high-speed rail network. Triple the length of the existing high-speed rail network by 2030 and maintain a dense railway network in all Member States. By 2050 the majority of medium-distance passenger transport should go by rail.” (EC, 2011: 9, own emphasis). The background for these bold objectives is the ambitious target with respect to reducing GHG emissions from transport and the relative environmental efficiency of rail on long distance journey compared to other modes (Givoni et al., 2009). “The Commission will ensure its actions increase the competitiveness of transport while delivering the minimum 60% reduction of GHG emissions from transport needed by 2050…” (EC, 2011:17).

While the majority of journeys are short in nature, the majority of GHG emissions from surface transport are the result of the long journeys (over 10 km). In the Netherlands about 33% of all trips are trips longer than 10 km and the share of these trips in total distances travelled is much larger: about 80%. The share of these trips in total emissions is even higher since many short distance trips are non-motorised trips (CBS, 2010). It is on these longer, surface transport journeys where rail has an important role to play in maintaining a competitive transport network which is also low in GHG emissions.

Railway stations are the access points to rail services and an important factor in determining the use of rail rather than other modes. In addition to the rail journey itself and its characteristics (speed, reliability, comfort, cost, etc.) the characteristics of the journey to get to and from the rail station are important, maybe even more in determining rail use (see Givoni and Rietveld, 2007, Brons et al., 2009). In principle, each additional rail station would increases rail’s attractiveness by increasing the accessibility of rail services. In most large cities there are several, often many, railway stations, making it an accessible mode of transport and easing the use of rail, overall increasing its competiveness compared to other modes. “Data show that 90% of Germans live within 1 km of public transport, compared to only 43% of Americans” (Buehler, 2011: 649) and this may partly explain the higher use of public transport in Germany compared with the USA (Buehler, 2011: 649). The implications of closing, rather than adding, a station can be deduced from Deka (2012) who examined the impacts of closing commuter rail stations to non-residents through parking restrictions. Implementing non-residents parking restrictions at railway stations clearly affect those residents, some are forced to travel longer distances to other stations (effect captured by the research) and some are probably discouraged from using rail at all (not captured by the research which was based on surveying rail passengers – Deka, 2012). But providing many stations comes at a cost, it slows the service for all those not using a particular station or stations and thereby making rail less attractive.

Examining the number of rail stations in a particular urban area has not received much attention to date, less for example than the attention given to planning bus networks and bus stops (e.g. Ibeas et al., 2010, Reinhold, 2008) although rail stations, their number and location, affect simultaneously the urban and inter-city transport networks and travel choices. Stations’ number and location is a key planning factor in planning both the inter-city and urban transport networks. Vuchic’s (1966) work can be considered the first to examine in detail the question of the number of stations, or station spacing in his terminology. When demand for rail travel started to increase after year of decline, for example in the UK since the beginning of the 1980s (see ATOC, 2008) research on this matter also increased. Blainey and Preston (2013) show how between 1970 and 1974 the net balance in the number of rail stations in the UK was 112 closed stations (18 new stations, 130 closed), while in each 5-year period that followed since up to 2009 the balance was always positive with a peak of 82 stations (96 new stations, 14 closed) between 1985 and 1989. Preston, 1991, Fowkes and Preston, 1991, Jones and White, 1994 and Wardman et al. (2007) all reflect the need of the time for improving rail demand models, part of which also concern demand for new rail stations and thus issues related to their number (all papers cited in Blainey and Preston, 2013). More recently and specifically addressing the question of opening a new rail station, Lebrun and Dobruszkes (2012) investigated the case for new stations on the Brussels RER network examining several alternatives for opening a new station by “calculating potential traffic, or a population basin” (p. 6). Blainey and Preston (2013) propose an appraisal framework for addressing the same question and apply it for a case of a new station in the rural area of South-east Wales. Another line of research, that addresses questions on number of rail stations, station location and thus account for access to rail stations, focuses more on the methodological aspects related largely to modelling the choice between stations – the key question addressed in this paper – and the use of modelling and forecasting tools for this purpose. The work by Hamacher et al. (2001), Schöbel (2005) and recently by Chakour and Eluru (2014) illustrates this line of research (Chakour and Eluru (2014) also review earlier research starting from the work by Liou and Talvitie, 1974).

The above brief review of the literature highlights a few shortcomings this research tries to address and illustrates in what this research differs from previous efforts to address questions related to the number of station. First, focus in past research has largely been on one station, even when several stations were considered there was no direct relation between them – the individuals stations were the concern not the set of stations. Second, the context and purpose of the research was forecasting. Third, the question has largely been if and where to open new stations. In the analysis described below, a set of stations that are currently in operation within one particular urban area is examined and the question asked is whether the number of stations ought to be changed. Given the current demand and its distribution between the stations is there room for considering opening new stations or there is room for closing a station or stations? To answer this question on the need for more or fewer station the logsum approach is adopted, a different optimisation approach then used before to address the same question. The focus of the analysis here is on the access journey, but the actual train journey and the destination are also accounted for in this analysis unlike in other research. In many ways, the research by Chakour and Eluru (2014) is similar in its settings to this research, but it is more concerned with the methodological rather than, as this paper is, the policy aspects of the subject. Still it provides important insights that will be referred to below.

To examine the number of railway stations in a multi-station urban area, and the implications it has for an optimal transport network, considering both the users and the suppliers, it is necessary to first understand how passengers choose a (departure) station when they have several options. In this context, and using the Amsterdam area as a case study, this paper aims to analyse the choice of a departure station by rail passengers and based on that examine the effect of closing a station on those passengers.

In the remainder of the paper we first summarise the findings from earlier research on access to rail in urban areas. We then present the data and methodology to examine the choice of a railway station (Section 3) followed by modelling results (Section 4). In Section 5 we examine the implications of closing a station in the Amsterdam area, which leads to a discussion on changing (reducing or increasing) the number of rail stations in Amsterdam. Finally, conclusions are presented (Section 6).