Conditional Elasticity of Rail Demand to Fares
Fitsum Teklu, SYSTRA, Mark Wardman, SYSTRA, Jeremy Toner, Institute for Transport Studies, University of Leeds
The paper discusses the methodology used in a recent research study into UK rail fare elasticities, extending Wardman and Toner (2003) Econometric Modelling of Competition between Rail Ticket Types.
Rail passengers typically have the option of choosing between different tickets for their journeys. In such an environment, estimating the demand for a particular ticket type requires an understanding of how the demand for that ticket is related to the demand and price of all other ticket types available, i.e. cross-elasticities. Estimating cross-elasticities is notoriously difficult, and even if they can be estimated, they are considered less transferable across time and space due to their dependence on specific market conditions. To avoid this, the UK rail industry primarily uses conditional elasticities, which measure the demand impact of increasing all prices by the same proportion. Market specific cross elasticities are derived from the conditional elasticities for use in demand forecasting.
This paper describes the methodology used in a recent research study into estimating conditional elasticities for the UK Rail Passenger Demand Forecasting Council. We present a brief overview of the research literature into estimating fare elasticities in such an environment. This is followed by a discussion of the econometric analysis undertaken using rail ticket sales data from 1994 to 2014, in which several approaches were used to constrain the relationships between the cross-elasticities for different tickets. The econometrics was complemented with a detailed market research. In the market research, passengers were presented with the alternative tickets available for their journey, and asked to choose a ticket a number of times after changing the fares. The market research yielded journey purpose specific cross- and own-elasticities, which were also scaled using the ticket sales data. The modelling extends the pioneering work by Wardman and Toner in the ETC paper of 2003.
Association for European Transport