The presentation is based on a paper co-authores by Loic Berger and Valentina Bosetti. This paper reports the results of an experiment jointly characterizing risk and model uncertainty aversion.
Based on a general framework developed in Marinacci (2015), we explore subjects’ attitudes towards risk (one stage objective uncertainty about the outcome of a given model), compound risk (two-stage objective uncertainty), and model uncertainty (subjective uncertainty about potential models generating the observations). We find that, provided the decision framework is numerically simple enough, subjects tend to be indifferent between risk and compound risk, while they exhibit larger aversion toward model uncertainty. This behavioral characteristics is in turn interpreted as evidence of ambiguity aversion. In addition, by fitting a flexibile utility model we quantify the degrees of aversion towards the different types of uncertainty, and find that, similarly to risk aversion, model uncertainty aversion is decreasing in absolute terms, while it increases in relative terms. These results have important implications for ambiguity aversion, which we show is decreasing when individuals become better off if considered in absolute terms, and constant if considered in relative terms.
This seminar has been jointly organized by FEEM and IEFE, Bocconi University.
Speaker:
Loic Berger, Federal Planning Bureau – Analyses et prévisions économiques, Belgium
Video-conference at FEEM Venice
Milano, FEEM - FEEM, Milano -
28 Jan 2016
Contacts Website Organized by
- FEEM - Fondazione Eni Enrico Mattei
- IEFE - Università Bocconi