Arrow's Theorem on the Optimality of Deductibles: A Stochastic Dominance Approach.
We provide a new proof for the optimality of deductible insurance that does not depend on the expected-utility hypothesis. Our model uses only first- and second-degree stochastic dominance arguments.