Weakening Description Invariance to Deal with Framing Effects: An Axiomatic Approach

The goal of this paper is to provide an axiomatic framework that can account for framing effects violating the so-called axiom of description invariance. Most existing discussions of the latter in economics are made with respect to Kahneman and Tversky's 1980s work. However, other psychologists have, over the last twenty-five years or so, progressively refined the conditions under which framing effects violating description invariance hold. Our axiomatic framework is motivated by these developments. We argue that description invariance is an implicit axiom of the standard model primarily on the formal structure of its object of choice and only derivatively on the formal structure of preferences. The conditions under which it is violated in psychologists' experiments are a useful guide to make it formally explicit. Furthermore, they also provide normative justifications for weakening it. We propose a way to do so in a mathematically tractable fashion that can account for all the variations in framing effects for which prospect theory cannot (at least straightforwardly) account for.