Non-expected utility and the robustness of the classical insurance paradigm

This paper uses the tools and techniques of generalized expected utility analysis to explore the robustness of some of the classical basic results in insurance theory to departures from the expected utility hypothesis on agents' risk preferences. The areas explored consist of individual demand for coinsurance and deductible insurance, the structure of Pareto-efficient bilateral insurance contracts, the structure of Pareto-efficient multilateral risk-sharing agreements, and self-insurance and self-protection. Most, though not all, of the basic results in this area are found to be quite robust to dropping the expected utility hypothesis.