The Welfare Evaluation of Risk Preferences
There are many settings in which policy-makers make welfare judgments about risky choices that people make. This workshop examines the economic foundations for these judgments, with an emphasis on measurement and quantification that is consistent with theoretical foundations. Is it possible for policy-makers to know when “mistakes” entail large welfare losses and when mistakes involve tiny welfare losses? Are there any useful tools to emerge from the field of behavioral welfare economics in terms of evaluating risk preferences?
The first day will focus on specific applications to gambling behavior, and how surveys and experiments can be used to evaluate the welfare costs of gambling. The relationship between welfare cost and the classification of gambling risk in terms of “problem gamblers” or “pathological gamblers” will be examined.
The second day will focus on broader methodological questions and different applications to insurance. Many behavioral economists characterize EUT as a normative theory of decision making under risk and non-EUT models as descriptive theories. This is loose: EUT is a descriptive theory that has some normative properties, whether or not it is a good descriptive theory. Conversely, some non-EUT models have known normative properties, apart from being allegedly better descriptive models. One example is the claimed “ecological rationality” of certain heuristics. Before we design policy to nudge someone in the right direction, what direction is “up”? Is it always EUT that provides the right metric, and if not what do we mean by the right metric for evaluating risk preferences? Similar questions might be asked about the normative status of Bayes Rule, which plays a critical role in the subjective risk perceptions that also affect behavior towards risk. And Bayes Rule is an additional piñata for constant flogging in Behavioral Finance.
This workshop will focus on longer presentations and discussions than usual, and more panel sessions. It is not intended to be a workshop that just focuses on the instant paper being presented, but at the broader questions posed above.