The hunt for a descriptive theory of choice under risk--A view from the road not taken

In this paper I propose that the development of descriptive theories of choice in economics has been profoundly influenced by an arbitrary and seemly innocuous decision as to how to present risky choices to experimental subjects. This decision to represent lotteries as prospects has lead to a preoccupation with the question of whether preferences conform to what is known as the "independence axiom." Had the profession chosen to represent lotteries in the action-by-state matrices favored by Savage, the independence axiom would have appeared uncontroversial but we would have questioned whether preferences obeyed arguably more fundamental tenets of rationality like transitivity. That different ways of representing lotteries lead to different conclusions regarding which axioms preferences do and don't obey suggests that the choices people make aren't necessarily reflecting properties of their preferences at all. Instead the choices reveal properties of the decision rule individuals use to try to satisfy their preferences - a rule that involves judgments regarding the similarity or dissimilarity of prizes and their associated payoffs across alternatives. The paper discusses how such judgments explain observed behaviors given both prospect and matrix representations of lottery choices as well as explaining anomalies in other choice domains.