The use of statistics for the management and regulation of insurance risks: the limits of the approach adopted by Solvency 2.

Summary The use of statistical magnitudes to inform decision-making in risk situations, which first appeared in the 18th century and was then disqualified in the 19th century, was reintroduced in the middle of the 20th century and has since then gradually gained acceptance in the financial industry, percolating through the insurance industry to the point where it has become widespread via Solvency 2. However, numerous dysfunctions linked to the mobilization of these tools for the management and regulation of risks have been highlighted by the economic, actuarial and management science literature. Consequently, this extension of the scope of application of such tools to the regulation of insurance raises questions. This thesis (i) empirically extends to the insurance sector the literature produced in other sectors in order to identify the limits of the use of statistics for risk management and prudential regulation, (ii) proposes a unified theoretical framework for the dysfunctions linked to these uses, and (iii) sheds light on the reasons for the adoption of these tools within the insurance sector.