Understanding the Demand for Rail Travel in UK

Class 395 "Javelin" high-speed train operated by Southeastern, at Ebbsfleet International

Class 395 "Javelin" high-speed train operated by Southeastern, at Ebbsfleet International

Photo by Sunil060902, CC BY-SA 3.0

Because predicted rail growth — based on ticket sales data — has not compared well to the observed increase in rail demand, the UK Department for Transport sought recommendations on how the current forecasting framework can be extended to incorporate external factors. An analysis of the National Travel Survey allowed researchers to quantify the impact of external socio-economic factors on rail demand.

Background

Over recent years, predicted rail growth has not compared well to the observed increase in rail demand. In the UK, the demand for rail travel is usually forecast using models based on ticket sales data to obtain the influence of service factors, such as rail fares, as well as exogenous factors, such as income growth.

Goals

The aim of the study was to better understand how factors external to the rail industry, like age, gender, employment patterns and car ownership, influence the demand for rail travel and to provide recommendations on how the current forecasting framework can be extended to incorporate these factors.

Methodology

The modelling approach used in the study combines the strengths of two data sources describing rail demand:

  1. disaggregate information on travel and travellers from the UK National Travel Survey (NTS) to quantify the impact of external socio-economic factors on rail demand, and
  2. aggregate time series ticket data to quantify the impact of income, rail service and the service levels of competing modes on rail demand.

RAND Europe’s work focussed on the analysis of the NTS data, which allows exploration of external factors like age, gender, employment type and car ownership on rail demand (as well as obtaining an independent set of income elasticities). To quantify the impact of external socio-economic factors on rail demand discrete choice models of rail trip making from (disaggregate) NTS data were developed. The models are structured to understand two issues related to rail demand: who travels by rail and how many trips rail users make.

Findings

  • Personal income is a strong determinant for the choice of using rail as a mode of travel. Across all journey purposes and geographies we observe that increasing income leads to an increase in the propensity to travel by rail, although increasing income levels do not seem to have such a large impact on the propensity to make multiple trips.
  • People with full driving licences are less likely to use rail for commuting journeys and other trips. Furthermore, as the number of cars in the household increases the propensity to travel by rail decreases. People who have a car freely available in the household are less likely to make rail trips.
  • The presence of a company car affects the propensity for rail travel differently for commuting and business travel. Commuters with company cars are less likely to make rail trips whereas business travellers with company cars are more likely to make rail trips (perhaps the presence of a company car is a proxy for job type). Overall, the trip rates for rail travel for business purposes are very similar for people with and without company cars in the household.
  • When commuting, full-time workers are most likely to make rail trips, followed by part-time workers, followed by those who are self-employed. Full-time workers are also more likely to make multiple rail commute trips.
  • For travel for other purposes, self-employed workers and temporarily sick people, disabled people and people looking after family are less likely to make rail trips relative to full time workers. In comparison, students, the retired, the unemployed and part-time workers are more likely to make rail trips for other purposes. Full-time workers are less likely to make multiple rail trips for other purposes.
  • Across all purposes, people working in managerial, professional or administrative occupations are more likely to travel by rail compared to those with other occupations.
  • In terms of employment type, those who work in the finance sector (for commuting and other travel) and real estate (for business travel) are more likely to travel by rail, whereas those who are involved in manufacturing, wholesale business, construction and health/social care sectors are less likely to travel by rail Moreover, for commuting, those who work in the financial sector are more likely to make multiple rail trips in the week for commuting purposes. Therefore, as the structure of the economy changes, we would expect changes in rail demand.
  • In general, older people and those under 16 years of age are less likely to travel by rail, whereas those who are employed and are under 25 years of age are more likely to make multiple rail commuting trips.

We observed a significant time-trend effects across most purposes and geographies, indicating an increased likelihood of travelling by rail over time that is not explained by socio-economic and network service terms. Some may hypothesise that this is a result of the digital revolution, which may have reduced the disutility of rail travel time both in absolute and relative to other modes

The changes in the employment mix of cities had been hypothesised to explain (some of) the strong growth in rail into Britain’s core cities; the NTS analysis has allowed us to quantify its effect so that it can be included both in our aggregate econometric models and in future forecasts.

Successful application of the improved models for forecasting purposes requires the collation of socio-economic and demographic forecast data at an appropriately granular level in order to capitalise on the framework developed in this study.