infonet-economy

Insurance contracts may fail to perform, leading to a total or partial default on valid claims. We extend models of such probabilistic insurance to allow for ambiguity in contract nonperformance risk, and derive formally that mean-preserving ambiguity reduces demand. The results of a field lab experiment are consistent with this logic. In particular, we find that a 10 percent contract nonperformance risk reduces insurance demand by 17.1 percentage points even when premiums are adjusted accordingly. Ambiguity about this contract nonperformance probability further decreases demand by 14.5 percentage points. While the demand-reducing effect of ambiguity is more pronounced for high-numeracy and ambiguity-averse individuals, it appears to be little affected by experience. The cause of an insurance contract failing to perform does not significantly influence the strength of these effects, but independently affects demand of low-numeracy and ambiguity-averse individuals.