Dual utilities on risk aggregation under dependence uncertainty (Q2274230)

The authors consider an application of Yaari's dual utility theory for decision analysis under uncertainty. A special case is the computation of the worst case utility of X+Y if the dependence structure of X and Y is not known. Lower and upper bounds on dual utilities are given. The main assumption for the existence of exact bounds is a generalisation of \textit{R. Wang} et al.'s definition `joint mixability' [Finance Stoch. 17, No. 2, 395--417 (2013; Zbl 1266.91038)] called conditional joint mixability (Theorem 4.3). The authors show that this condition is necessary for special distortion functions. Some economic applications are given: A decision maker equipped with a dual utility tends to be more pessimistic about the aggregate risk. It should be noted that dual utility -- like expected utility theory -- has some drawbacks concerning behavioral patterns.