Age, income, and the discounting of delayed and probabilistic rewards

Recent meta-analyses of age-related changes in decision-making have yielded inconsistent findings, reporting no sizable age effect on intertemporal and risky choice (i.e., delay and probability discounting, respectively). However, these reviews did not systematically account for socioeconomic status, a potential moderator. The present research addresses this gap by examining the interactive effects of age and self-reported annual household income on both intertemporal and risky choice across a wide range of reward amounts. In two parallel studies, participants aged 20 to 80 (Study 1, delay discounting: N = 596; Study 2, probability discounting: N = 592) completed an adjusting-amount discounting task with hypothetical rewards of $150, $2,500, and $30,000. As expected, income was unrelated to discounting in participants under 35. For those 35 and older, the effects of age were domain-specific: Degree of delay discounting decreased with age, as indicated by linear increases in the Area-under-the-Curve (AuC), whereas for risky choice, the relation between AuC and age was nonlinear, with discounting decreasing until around age 60 and relatively little change thereafter. As predicted by the buffering hypothesis, these age effects were moderated by income in both domains. For delay discounting, the age-related decrease in degree of discounting was present only in lower- and middle-income groups. For probability discounting, the nonlinear age trajectory was absent in the high-income group. These findings resolve prior inconsistencies by demonstrating that the effects of aging on economic preferences follow distinct, domain-specific trajectories that are markedly influenced by socioeconomic context.