Cognitive decline is associated with risk aversion and temporal discounting in older adults without dementia

Abstract

Risk aversion and temporal discounting are preferences that are strongly linked to sub-optimal financial and health decision making ability. Prior studies have shown they differ by age and cognitive ability, but it remains unclear whether differences are due to age-related cognitive decline or lower cognitive abilities over the life span. We tested the hypothesis that cognitive decline is associated with higher risk aversion and temporal discounting in 455 older persons without dementia from the Memory and Aging Project, a longitudinal cohort study of aging in Chicago. All underwent repeated annual cognitive evaluations using a detailed battery including 19 tests. Risk aversion was measured using standard behavioral economics questions: participants were asked to choose between a certain monetary payment versus a gamble in which they could gain more or nothing; potential gamble gains varied across questions. Temporal discounting: participants were asked to choose between an immediate, smaller payment and a delayed, larger one; two sets of questions addressed small and large stakes based on payment amount. Regression analyses were used to examine whether prior rate of cognitive decline predicted level of risk aversion and temporal discounting, controlling for age, sex, and education. Over an average of 5.5 (SD=2.9) years, cognition declined at an average of 0.016 units per year (SD=0.03). More rapid cognitive decline predicted higher levels of risk aversion (p=0.002) and temporal discounting (small stakes: p=0.01, high stakes: p=0.006). Further, associations between cognitive decline and risk aversion (p=0.015) and large stakes temporal discounting (p=0.026) persisted in analyses restricted to persons without any cognitive impairment (i.e., no dementia or mild cognitive impairment); the association of cognitive decline and small stakes temporal discounting was no longer statistically significant (p=0.078). These findings are consistent with the hypothesis that subtle age-related changes in cognition can detrimentally affect individual preferences that are critical for maintaining health and well being.

Fig 1. Association of cognitive decline with risk aversion as derived from a non-linear mixed effects model.

The figure depicts the probability of taking the gamble as a function of the gamble gain in dollars. Lower curves on the Y axis indicate more risk aversion. Predicted curves are shown for a typical participant (i.e., female with median age, education, and income) at three different levels of cognitive decline: red indicates fast (highest 10^th percentile), blue indicates median, and green indicates slow (lowest 10^th percentile) cognitive decline.

Fig 2. Association of cognitive decline with temporal discounting (small and large stakes) as derived from a non-linear mixed effects model.

The figure depicts the probability of taking future payment rather than a fixed immediate payment as a function of the future payment in dollars. Lower curves on the Y axis indicate more temporal discounting. Predicted curves are shown for a typical participant (i.e., female with median age, education, and income) at three different levels of cognitive decline: red indicates fast (highest 10^th percentile), blue indicates median, and low indicates slow (lowest 10^th percentile) cognitive decline.