Using hypothetical lottery choices to measure risk preferences, Frederick (2005) finds that higher cognitive ability is associated with less risk aversion. This paper documents, however, that when using an incentive compatible measure of risk preference, attitudes towards risk are not associated to cognitive ability as measured by Frederick's (2005) three-item cognitive reflection test. This is a new finding that adds weight to the claim that lack of proper financial incentives can sometimes be a source of bias. In addition, we show that this lack of association between risk preferences and cognitive ability is robust to using a broader measure of cognitive ability that takes into account both verbal and non-verbal reasoning skills. Our results suggest the possibility that whether cognitive ability relates to attitudes towards risk is sensitive to instruments used to measure both of them.
cognitive ability risk preferences financial incentives cognitive reflection test