On Two Points of V Iew regarding Revealed Preference and Behavioral Economics

The concept of revealed preferences surfaces quite often in methodological debates over behavioral economics. There seems to be a close linkage between economists’ attitudes to the revealed preference principle and the opinions that they hold regarding paternalistic policies, or their relative evaluation of decision models. This chapter is an attempt to clarify several aspects of this linkage. The revealed preference principle is part of a philosophical tradition that places restrictions on professional discourse, by judging parts of it to be “meaningless.” It was originally formulated at a time when economic models involved little more than static choice of a consumption bundle from a budget set. In this simple consumertheoretic environment, a single “mental construct” is attributed to the decision maker, namely, a utility function u : X → R. The revealed preference principle then means that any property of u is meaningful only if it can be defined in terms of a preference relation R ⊆ X × X , which in turn can be elicited from choice experiments whose outcomes are summarized by a choice correspondence c , where c(A) ⊆ A for every nonempty A ⊆ X . To quote Gul and Pesendorfer (chapter 1):