Safety First and Ambiguity

There is considerable empirical evidence suggesting that ambiguity (i.e., parameter risk) impacts pricing decisions by actuaries and underwriters and their desire to provide coverage. Stone proposed a safety first model of choice that provides a possible explanation for this behavior. This paper analyzes Stone's proposed stability and survival constraints and compares the results with those predicted by expected utility theory. The analysis is motivated by insurers' increaSing reluctance to provide coverage for certain specific risks such as earthquake damage insurance where the probability of loss is ambiguous. We show that such behavior is consistent with safety first but is difficult to explain using an expected utility approach.