Dynamic analysis and forecasting of public transport demand using a combination of demographic and econometric approaches

Document Type

Conference Paper

Publication Date


Subject Area

ridership - forecasting, ridership - forecasting, ridership - demand


Public transport use in France has undergone two distinct periods of change in the last twenty years. A period of sustained increase in public transport traffic linked with favourable policies (in the 1970s the employers' payroll tax for transport was created and the domestic transport policy act was passed in 1982, etc.) was followed, from the second half of the 1980s, by a period of difficulties and loss of market share for public transport (the effects of urban sprawl and tt~e spreading out of flows) [Bonnel 2000]. Two types of approach are adopted to investigate the factors which influence the use of urban public transport (UPT) and its change. Demographic studies have demonstrated the extent to which public transport use is sensitive to the level of household car ownership [Webster et al. 1985] and to type of home location and age. A study by [Massot 1991] .added the quality of public transport supply to these criteria and showed that sensitivity to supply could vary between different groups within the population. Due to a lack of data, it has however not been possible to validate these analyses in a changing context. Taking a dynamic approach, age cohort models [Boulahbal and Madre, 1999] have proved to be an effective means of investigating long term changes in household car ownership, but it is difficult to apply such models to travel as they cannot take into account changes in the supply and pricing context. On the other hand, econometric studies have been used to model change in UPT use and estimate its elasticity with respect to transport supply and price. Such studies are frequently experimental and short term and are based on aggregate indicators which either ignore changes in population structure or do not consider them in sufficient detail 1. VIA GTI, INRETS and ERMES have performed a retrospective and foreward study of UPT use which attempts to estimate both the effects of population structure (demographic approach) and the effects of supply and price (econometric approach). This study analyzes a panel of 61 provincial Urban Transport Areas 2 existing since 1975. This paper presents the methodology used for this study and its principle findings. It is in two parts, the first describes the methodology by which econometric modelling takes account of population structure, and the second describes the estimated econometric model and how this has been exploited.