Cognitive abilities and risk-taking: Errors, not preferences

Investigating inconsistencies in complex lotteries: The role of cognitive skills of low-numeracy subjects

“…Cognitive skills do not remain a statistically significant driver when only considering monotonous choice patterns. This finding is in line with Amador-Hidalgo et al (2021), who find no effect of cognitive skills on inattentive behavior among their comparably highly educated sample of students.…”

“…Taking education as a proxy for cognitive skills, they find no relationship and suggest that inconsistency levels rise due to measuring ambiguity rather than risk attitude and put forward that they "(...) leave it for future research to formally test for cognitive skills as a potential underlying reason of the inconsistency (...)" (He et al (2018(He et al ( ), p.1968. While this relationship between cognitive skills and errors within complex lotteries has been investigated in western societies (see for example Amador-Hidalgo et al (2021); Andersson et al (2016)), to the best of our knowledge there is no evidence stemming from low-numaracy subjects, especially not rural farmers. This is the starting point of our study: We hypothesize that cognitive skills -measured through the RPM test -are a negative and statistically significant determinant of inconsistency levels in the HL-task for low-numeracy subject pools .…”

“…Second, a strict indicator (Inconsistency HL-task II), where we add those who chose the dominated option in row 10, i.e. indicating the lack of attention rather than a calculation error (Amador-Hidalgo et al, 2021). In other words, Inconsistency HL-task II captures all participants that violated first-order stochastic dominance in the lottery.…”

Comparing Risk Elicitation in Lotteries with Visual Or Contextual Framing Aids

“…This conclusion is also supported by previous research on the negative relationship between cognitive abilities and risk aversion. If treatment BB provides more clarity, and subjects consider the task less complex than HL, this may be equivalent to subjects having more cognitive ability, therefore making fewer errors (less inconsistencies) and showing less risk aversion (Amador-Hidalgo et al, 2021). Our results may be particularly relevant for risk elicitation experiments in developing countries, where the percentage of inconsistencies is usually high.…”

“…The reduction in the elicited risk aversion coefficient in treatment BB can be interpreted using the results of Amador-Hidalgo et al (2021). They find that low cognitive ability subjects face a higher computational complexity (and choose randomly) in Holt and Laury (2002) task after some point, precisely when consistent individuals start choosing the risky option more often.…”