Is Expected Utility an Ex-Hypothesis? Some Implications of Reference-Dependent Expected Utility Model

Abstract

Rabin and Thaler (2001) declared Expected Utility an ex-hypothesis or a dead parrot alluding to the famous sketch from Monthy Pythons Flying Circus. Following Cox and Sadiraj (2006) and others, one should distinguish between Expected Utility (EU) theory (a purely mathematical theory based on axioms) and Expected Utility models (EU theory plus a given economic interpretation). The most prevalent EU model is one that assumes consequentialism (Rubinstein, 2012). Consequentialism states that the decision maker has a single binary preference relation comparing probability distributions over final wealth levels. Preference relations over wealth changes for different levels of wealth are derived from this single preference relation. EU theory plus consequentialism is referred to as the standard EU model. It is argued that most of the critique against EU is against the standard EU model, or against consequentialism. We replace consequentialism with reference-dependence, retaining the EU hypothesis. Using Sugden (2003) framework, we show that many violations of the standard EU model can be explained assuming this different interpretation. Among the topics considered are: WTA/WTP disparity, preference reversal, complementary symmetry, preference homogeneity, loss aversion, reflection effect and the coexistence of insurance and gambling.