Modelling the Impact of Alternative Fare Structures on Train Overcrowding

Modelling the Impact of Alternative Fare Structures on Train Overcrowding


G Whelan, D Johnson, ITS, University of Leeds, UK



The PRAISE (Privatisation of Rail Services) model was initially developed at the Institute for Transport Studies to look at the potential for open access competition on the Leeds?London corridor but it has subsequently been applied to other routes in the UK (e.g. Gatwick Express) and overseas (e.g. Stockholm?Gothenburg). More recently, the model has been re-written and developed for the Strategic Rail Authority (SRA) as a Windows software package capable of assessing many routes and operating conditions.

The software comprises a demand model, a cost model and an evaluation model. The demand model has a hierarchical structure and works at the level of the individual traveller. Using information on passenger?s valuation of journey attributes, such as journey time and service frequency, together with elasticity estimates, the lower level of the model assigns a probability that a given traveller will choose a particular ticket and outward and return service combination. By aggregating the ticket and service probabilities over a representative set of simulated passengers, the model is able to forecast market shares for each service and ticket combination. To allow for the fact that changing fares and services will change the overall demand for rail, the upper level of the model is structured to allow for the rail market to expand or contract according to the overall level of service. By assessing the outward and return portions of a journey, together with information on ticketing restrictions (departure time, advanced purchase, transferability between operators) the model is able to forecast ticket revenue by operator.

The cost model employs an accountancy cost approach incorporating: costs that are related to operating hours, costs that are related to train kilometres and fixed costs.

Costs can be varied by operator and rolling stock type and can be combined with estimates of revenue to generate forecasts of operator profitability.

The model generates output that can be used in a formal appraisal system. This output includes, passenger demand, passenger distance, operator revenue, operator costs, profitability, user benefits (consumer surplus), overcrowding, and diversion to and from other modes in terms of passenger numbers and passenger distance.

The SRA provides the backbone to rail regulation in Great Britain. As part of its responsibilities, the SRA monitors overcrowding on trains which it measures in terms of the proportion of passengers on trains in excess of the seat capacity for longer distance services, and with an allowance for standing passengers on shorter journeys of less than 20 minutes. Overcrowding on Britain?s railways fell during the early 1990s but as been on the increase since 1996 with particularly acute problems in the morning peak for services travelling to London.

In an interesting study conducted on behalf of the SRA we adapted the PRAISE model to include generalised journey cost penalties for overcrowding based upon journey purpose, journey time and degree of overcrowding. Using demand, fares and timetable information for an actual case study route, we examine how fares and ticketing restrictions can be set to manage demand throughout the day without significantly reducing the overall demand for rail travel. The implied cross elasticities between ticket types are then compared against recent empirical findings discussed in Wardman and Toner (2003).


Association for European Transport