Self-Confidence, Overconfidence and Prenatal Testosterone Exposure: Evidence from the Lab

This paper examines whether the degree of confidence and overconfidence in one's ability is determined biologically. In articular, we study whether foetal testosterone exposure correlates with an incentive-compatible measure of confidence within an experimental setting. We find that men (rather than women) who were exposed to high testosterone levels in their mother's womb are less likely to overestimate their actual performance, which in turn helps them to gain higher monetary rewards. Men exposed to low prenatal testosterone levels, instead, set unrealistically high expectations which results in self-defeating behaviour. These results from the lab are able to reconcile hitherto disconnected evidence from the field, by providing a link between traders'overconfidence bias, long-term financial returns and prenatal testosterone exposure.

Choice, Deferral and Consistency

In this paper we study decision making in situations where the individual’s preferences are not assumed to be complete. First, we identify conditions that are necessary and sufficient for choice behavior in general domains to be consistent with maximization of a possibly incomplete preference relation. In this model of maximally dominant choice, the agent defers/avoids choosing at those and only those menus where a most preferred option does not exist. This allows for simple explanations of conflict-induced deferral and choice overload. It also suggests a criterion for distinguishing between indifference and incomparability based on observable data. A simple extension of this model also incorporates decision costs and provides a theoretical framework that is compatible with the experimental design that we propose to elicit possibly incomplete preferences in the lab. The design builds on the introduction of monetary costs that induce choice of a most preferred feasible option if one exists and deferral otherwise. Based on this design we found evidence suggesting that a quarter of the subjects in our study had incomplete preferences, and that these made significantly more consistent choices than a group of subjects who were forced to choose. The latter effect, however, is mitigated once data on indifferences are accounted for.