Contract Nonperformance Risk and Ambiguity in Insurance Markets

Insurance is referred to as probabilistic when policyholders are exposed to contract nonperformance risk, that is, claims are not reimbursed by the insurer. We extend probabilistic insurance models to allow for ambiguity regarding contract nonperformance and loss probabilities. We empirically test theoretical predictions from our model within a field lab experiment in a low-income setting. This is a persuasive context, since especially in emerging and poorly regulated markets there is significant contract nonperformance risk. In line with our predictions, insurance demand decreases by 17 percentage points in the presence of contract nonperformance risk and by 32 percentage points when contract nonperformance risk is ambiguous. Furthermore, ambiguity does not easily disappear with experience. The results have implications for both industrialized and developing insurance markets.