Eliciting Utility for (Non)Expected Utility Preferences Using Invariance Transformations

This paper presents a methodology to determine the preferences of an individual facing risk in the framework of (non)-expected utility theory. When individual preference satisfies a given invariance property, his utility function is solution of a functional equation associated to a specific transformation. Conversely, there exist transformations characterizing any given utility function and its invariance property. More precisely, invariance with respect to two transformations uniquely determines the individual utility function. We provide examples of such transformations for CARA or CRRA utility, but also with any other utility specification and discuss the example of DARA and IRRA specifications.