A Theory of Coarse Utility

Abstract

This article presents a descriptive theory for complex choice problems. In line with the bounded rationality assumption, we hypothesize that decision makers modify a complex choice into some coarse approximations, each of which is a binary lottery. We define the value of a best coarse approximation to be the utility of the choice. Using this paradigm, we axiomatize and justify a new utility function called the coarse utility function. We show that the coarse utility function approximates the rank- and sign-dependent utility function. It satisfies dominance but admits violations of independence. It reduces judgmental load and allows flexible judgmental information. It accommodates phenomena associated with probability distortions and provides a better resolution to the St. Petersburg paradox than the expected and rank-dependent theories.