The sensitivity of rail demand to variations in motoring costs: Findings from a comparison of methods

Abstract

Listen

Translate article

There is not a great deal of reliable summary evidence on how variations in motoring costs impact on the demand for rail travel, in contrast with evidence on other key demand parameters. In part this is because cross-elasticities are difficult to estimate whilst we must recognise that they are strongly dependent upon relevant market shares. In addition, the combination of car being the main competitor to rail and anticipated significant reductions in car costs over time mean that this is an important area for research. Given the historic challenges of estimating robust crosselasticities, the research reported here used a variety of methods based upon purpose collected survey data, large data sets of tickets sales and National Travel Survey data, major reviews of existing evidence, and deducing crosselasticities using economic theory. The findings provide a significant contribution to understanding in this area. There is a high degree of similarity in the fuel price cross-elasticities obtain by the different methods on average, although this masks variations across contexts. The cross-elasticities tend to be relatively small, which is in line with the survey evidence that not all motorists consider even fuel costs or are aware of changes in them. Nonetheless, crosselasticities can be large where rail is in a weak competitive position. Our findings challenge current conventions in the railway industry in Great Britain, as contained in the longstanding Passenger Demand Forecasting Handbook (PDFH), both in terms of the magnitude of cross-elasticities used and their variation across routes. Our preference is to deduce fuel price cross-elasticities from car own-price elasticities using diversion factors and relative demand of car and rail. This method is able to produce context specific cross-elasticities, reflecting the inherent variability of such parameters, whilst the econometric analysis and other evidence support the use of this approach. The method can be easily extended to cover parking costs, which we here did to provide new guidance for PDFH, and car journey times. We recommend means by which the method can be enhanced.