Maxmin Expected Utility over Savage Acts with a Set of Priors

Abstract: This paper provides an axiomatic foundation for a maxmin expected utility over a set of priors (MMEU) decision rule in an environment where the elements of choice are Savage acts. This characterization complements the original axiomatizations of MMEU developed in a lottery-acts (or Anscombe-Aumann) framework by Gilboa and Schmeidler (1989). MMEU preferences are of interest primarily because they provide a natural and tractable way of modeling decision makers who display an aversion to uncertainty or ambiguity. The novel axioms are formulated using standard sequence techniques, which allow cardinal properties of utility be expressed directly through preferences.