A life Expectancy-Period-Cohort model to project private car fleet and traffic applied to France

• Published in: Transportation research procedia (Online) 2020, Vol.48, p.1526-1545
• Main Authors: Madre, Jean-Loup, Bussière, Yves D., Grimal, Richard
• Format: Article
• Language: English
• Subjects: demographic; economic factors; Humanities and Social Sciences; Individual motorization; life expectancy; prospective; Sociology; use of the automobile
• ISSN: 2352-1465; 2352-1465
• DOI: 10.1016/j.trpro.2020.08.196

In most industrialized countries, after decades of gradually slowed growth, car traffic stagnated in the 2000s. This phenomenon has been attributed not only to conventional economic factors (stagnation of incomes, upward volatility in fuel prices) and to re-urbanization linked to metropolisation, but also to demographic factors (ageing of the population, longer life cycle stages leading in particular to delay the passage of the driving license in the younger generations). The economic recovery, albeit rather slow, and a significant drop in the price of oil in 2014 favored a certain revival of traffic growth in several countries (U.S.A., Germany, France, ...); but what about the structural factors and how to predict medium-term developments? We have already dealt with these questions via Age-Period-Cohort models, and more often Age-Cohort (AC). In view of the over-determination generated by the mechanical link between these three factors, we propose a Life Expectancy-Period-Cohort model (EPC); indeed, by replacing age by life expectancy at this age and at each date, the model can be directly estimated keeping the three components, while making this approach more consistent with the extension of life cycle stages (longer studies, women having their children in their thirties, postponement of retirement age, ...). Period effects are specified by introducing the income of the household and a fuel price index as explanatory variables. The results are compared with those of various previous models. The scope is the adult population (i.e., of driving age), considering three phases for automobile behavior:- to pass the driver’s license,- to be the main user of a vehicle,- to ride (annual mileage) or frequency of use of the vehicle. Once the model is estimated on the data of the Parc-Auto Kantar-SOFRES 1994-2016 panel survey, an example of medium-term (horizon 2030) projection of the annual mileage is presented, being aware that in the long term the technical innovations (autonomous vehicle, electric and hybrid engines) and organizational evolution (car sharing, carpooling, ...) are likely to fundamentally change the conditions of use of the car.