A Comparison of Survey and Incentivized-Based Risk Attitude Elicitation

This paper reports results from an on-line economics experiment with heads of households that explores the link between a sample of survey questions on the Canadian Know-Your-Client survey form and several incentivized laboratory instruments, both aimed at measuring risk attitudes and loss aversion. We find that the instruments significantly predict responses to risk questions, with the exception of a question that includes a time dimension. By contrast, the loss aversion instruments do not predict responses to loss questions. Indeed, if anything, risk instruments predict the majority of the loss questions. We conclude that the survey appears to be successful in eliciting attitudes towards risk; that the survey appears to be less successful with regard to loss aversion; and that it may be useful to include survey questions about higher order risk preferences on the form.