Abstract
We examine the relationship between confidence in own absolute performance and risk attitudes using two confidence elicitation procedures: self-reported (non-incentivised) confidence and an incentivised procedure that elicits the certainty equivalent of a bet based on performance. The former procedure reproduces the “hard-easy effect” (underconfidence in easy tasks and overconfidence in hard tasks) found in a large number of studies using non-incentivised self-reports. The latter procedure produces general underconfidence, which is significantly reduced, but not eliminated when we filter out the effects of risk attitudes. Finally, we find that self-reported confidence correlates significantly with features of individual risk attitudes including parameters of individual probability weighting.
Highlights
In this paper we report an experiment investigating relationships between measures of individuals’ confidence assessments of their own performance and their risk attitudes.Our broad motivation flows from a large literature originating in psychology in the 1970s and documenting apparently systematic biases in individuals’ assessments of their own abilities, both relative to others and in absolute terms
Because we are interested in possible relationships between individuals’ confidence judgements and their risk attitudes, which we interpret as features of individuals, we focus mainly on evidence related to the calibration of own absolute performance
The 45-degree line provides a natural benchmark in the sense that a general tendency towards overconfidence would result in points located above the line whereas a general tendency towards underconfidence would result in points below it
Summary
In this paper we report an experiment investigating relationships between measures of individuals’ confidence assessments of their own performance and their risk attitudes.Our broad motivation flows from a large literature originating in psychology in the 1970s and documenting apparently systematic biases in individuals’ assessments of their own abilities, both relative to others and in absolute terms. A range of studies using approaches like this, starting with the classic study of Fischhoff et al (1977), document systematic miscalibration, usually in the form of either overconfidence (i.e., over-predicting own actual success rate) or a hard-easy effect (i.e., overestimating success for ‘hard’ tasks and underestimating success for ‘easy’ ones).1 This literature has, in turn, stimulated significant streams of work in both empirical and, more recently, theoretical economics. One of the first papers in experimental economics to study absolute confidence miscalibration is by Blavatskyy (2009) In his experiment, subjects answer a set of 10 multiple choice quiz questions before choosing between two payment schemes. By contrast and as explained below, we do find significant correlations between risk attitudes and confidence.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have