This thesis aims to better understand the individual decisions of insurance and reinsurance market actors facing different information types, the coverage of extreme risks, and the contractual arrangements between insurers and reinsurers. The insurance context of extreme risks is characterized by the intensification and the increased number of natural and industrial disasters for the last twenty years in the world, which makes it difficult to properly evaluate the event characteristics and questions the insurability of these risks due to the presence of uncertainties. The thesis is composed of two parts using a wide range of quantitative tools. The first part focuses on the analysis of information effects on insurance decisions, especially the premium that actors are willing to pay or to accept in order to transfer or to take specific risks. The available information can lead to situations of risk or ambiguity, distinguishing two ambiguity types: imprecision in which the information is consensual but imprecise, and conflict where there is a disagreement between experts. This part is based on surveys distributed to a representative sample of the population as potential insurance buyers (chapter 2) or to insurance professionals (chapter 3) in order to study demand and supply behaviors. This part also presents an experimental approach in which participants are paid according to their performance (chapter 4). The second part examines the reinsurance market. After presenting the characteristics and mechanisms of the market (chapter 5), reinsurance supply is examined through auction behaviors of reinsurers (chapter 6). 8ased on a proprietary data-base, this part identifies the underlying factors of reinsurance premiums which depend on the risk characteristics, but also on the treaties, the reinsurers and the market.
