Contributions to risk assessment in storm and automobile insurance.

Summary In this thesis, we study storm coverage dedicated to wind damage and a development of behavioral insurance through automobile risk. We combine external information such as wind speed with insurance data. We propose the construction of a storm index to complete and reinforce the evaluation of damages caused by major storms. We then define a division of the French territory into 6 storm zones, depending on the extreme wind correlations, to test several scenarios. These different tests and considerations allow us to improve our storm index. We use extreme value theory models to show the impact of variability on the calculation of return periods and capital requirements. We thus highlight the difficulties encountered in obtaining robust results in relation to extreme events. In the case of motor insurance, we test different methods to respond to technical and regulatory changes. We characterize the male/female partition using the logistic procedure, multiple correspondence analysis or classification trees. We show that it is possible to compensate for the absence of the gender variable by other information specific to the insured or to his vehicle and in particular the use of mileage records. Finally, we are interested in the experience acquired by novice drivers. We study the on-road behavior of the insured to create new risk classes.