Mistakes in Choice-Based Welfare Analysis

Economics has always been concerned not only with describing or predicting economic behavior but also with understanding economic well-being. The traditional, official way economists (claimed to) have assessed well-being is from " revealed preference''—observing what people choose under the maintained hypothesis of 100 percent rationality. For instance, when we teach millions of students each year the conditions under which interfering with free-market exchange will make people worse off, by interfering with satisfying their wants, we do so under the compelling assumption that people tend to make choices that rationally maximize their own well-being. While economists are humble about the fact that we are not in a privileged position to declare what goals society should pursue, welfare economics has provided guidance on the determinants of well-being according to this restrictive set of criteria. Although all of us also assess well-being in other ways, it is only recently that economists have begun to do so in a more focused way. A growing number of researchers have begun to study alternatives to the 100 percent rationality assumption, while other researchers have used sundry techniques to measure well-being directly. In this article, we explore some conceptual issues as to why and how one might use these new assumptions and approaches to supplement and modify the revealed-preference approach as it is conventionally conceived. We take the central question of welfare economics to be: how do different situations or economic environments affect people's well-being? When doing so with sensible ancillary assumptions , inferring people's well-being from their choices, based on the presumption that they are rationally pursuing their goals, is, in our view, the best scientific approach to research on well-being yet formulated. Our first goal in this paper is to clarify just how crucial these ancillary assumptions are to rational-choice welfare analysis. Whether unnoticed or unemphasized, assumptions that are unobservable in choice behavior drive all welfare conclusions in economics. Any combination of observed behavior and assertions about what environments enhance well-being is consistent with utility maximization. The basic logic, formalized as a mathematically trivial theorem in Ko ˝szegi and Rabin (2007), is simple. As is clear from psychology , and recently has been better appreciated and elegantly modeled within economics, well-being may depend not only on the outcome resulting from choice, but also on the choice set itself. The effect of different choice sets on well-being is not observable by the choices made within each …