Optimism and Self-Selection Models: A Theoretical Investigation of the Rothschild-Stiglitz Insurance Model*

This paper analyzes the impact of biased beliefs on the structure of informational equilibria in classical self-selection models. Using the intuitive insurance market example of Rothschild-Stiglitz (1976) I show that biases of high-risk individuals have fundamentally different effects on equilibrium contracts than biases of low-risk individuals. Whereas equilibrium contracts of both groups are affected by biases of the high-risk group, equilibrium contracts are robust to moderate biases among the low-risk group. Specifically, optimism of the high-risk group results in underinsurance of their own group and causes negative externalities on the low-risk group by restricting the set of feasible contracts. If optimism is sufficiently strong, it is possible, that a breakdown of the insurance market occurs. This extreme result is more likely to have empirical relevance in insurance markets where losses are relatively small. In addition, I reveal that my analysis is not limited to the insurance market context by outlining a simple education screening model in the spirit of Spence (1972) that produces similar results.