The aim of this PhD thesis is to analyze the volume and the demographic structure of the French passenger car fleet in order to estimate the contribution of the car traffic to CO2 emissions and of local pollutants and to make related statistical projections to 2020. We used 'Parc-Auto' an annual car fleet database built from annual postal surveys conducted with a panel of 10 000 French households (annual observations over the period 1984-2008). The continuous increase of the average age shows that the passenger car fleet undergoes an ageing of structural order. The calibration of a family of parametric survival functions shows a continuous increase of vehicles life expectancy. The median life expectancy estimated by adjustment of a Weibull law on transversal survival rates increased from 8,6 years in the period 1984-1987 to 13,2 years in the period 2004-2008 which represent a gap of 4,6 years over the whole period. The increase of the life expectancy of vehicles is explained by households' multi-equipment, the increasing diesel share and by a relative increase of vehicles robustness. This relative increase of passengers cars lifetime added to the relative stability of new registrations observed from two decades explains the natural balance growth as well as the national fleet ageing. Moreover, a Poisson regression analysis allowed us to confirm the superiority of the diesel motor vehicles as well as the negative effect of the annual mobility on vehicles lifetime. Finally, on the basis of our survival study a projection to 2020 of the fleet structure following two different scenarios of hybrid and electric vehicles market shares has been conducted. For each fleet structure we obtained projections of CO2 and pollutants emissions until 2020 thanks to the software Copert IV.
