In this paper, household annual car mileage is analysed in 10 French zones, defined by crossing the criterion of distance to centre with that of conurbation population size, of which one groups rural areas. Using data from repeated cross-sections of the annual INSEE Household Conjuncture Survey (1977-94), pseudo-panels are constructed according to the birth year of the household head. Applying an Age-Cohort-Period model, estimates are made for age and cohort effects, along with income and price effects reflecting the general economic context faced by households during the period considered. This approach is advantageous in that it avoids imposing a priori saturation levels, it accounts for both economic and demographic factors (thus accounting for changes in the population structure and for the replacement of generations), and it uses reliable variables (demographic projections) for long term forecasts. Relying on the estimated effects and on demographic projections (number of households by age of the head in each zone), projections were undertaken of the car traffic generated by households for the years 2005, 2010, 2015 and 2020 according to three scenarios of urban sprawl, in addition to scenarios of growth of consumption and fuel prices. The results of the 1999 census in France confirm the tendency to urban sprawl so, a (High) scenario extending the tendency to urban sprawl observed from 1975 to 1990, an (Intermediate) scenario accounting for the inflexion observed between the censuses of 1990 and 1999, and a (Low) scenario fixing the spatial distribution of the population at its observed situation at the 1999 census were investigated. The gap between the two extreme scenarios widened with time. Though it remains modest at the national level (4% in 2020), this gap increases as the conurbation size decreases. The growth of car traffic between 1990 and 2020 should be larger the smaller the population size of the conurbation. The 5-year growth rates are lower the farther the horizon of projection in each of the scenarios, and their levels diminish as one moves from the High scenario to the Low one. Globally, the traffic generated by inhabitants of city centres should be more important in the case of a stop of urban sprawl (Low scenario). On the contrary, the traffic generated by inhabitants of inner and outer suburbs should be more important if urban sprawl followed its past tendency (High scenario). As regards the economic scenarios, the projections are more sensitive to variations in household final consumption than in fuel prices. For the covering abstract see ITRD E126595.
Abstract