Prospect Theory Versus Expected Utility Theory Assumptions, Predictions, Intuition and Modelling of Risk Attitudes

Abstract

The main focus of this tutorial/review is on presenting Prospect Theory
in the context of the still ongoing debate between the behavioral (mainly
descriptive) and the classical (mainly normative) approach in decision theory
under risk and uncertainty. The goal is to discuss Prospect Theory vs. Expected
Utility in a comparative way. We discuss: a) which assumptions (implicit and
explicit) of the classical theory are being questioned in Prospect Theory; b)
how does the theory incorporate robust experimental evidence, striving, at the
same time, to find the right balance between the basic rationality postulates of
Expected Utility (e.g. monotonicity wrt. First-Order Stochastic Dominance),
psychological plausibility and mathematical elegance; c) how are risk attitudes
modeled in the theory. In particular we discuss prospect stochastic dominance
and the three-pillar structure of modeling risk attitudes in Prospect Theory
involving: the non-additive decision weights with lower and upper subadditivity
and their relationship to the notions of pessimism and optimism, as well as
preferences towards consequences separated into preferences within and across
the domains of gains and losses (corresponding to basic utility and loss aversion),
d) example applications of Prospect Theory.