20-Year income volatility and cognitive function in the National Longitudinal Survey of Youth (1979) cohort: A replication and extension of CARDIA findings

Abstract

Grasset et al. previously reported that income volatility throughout early-to-mid adulthood (occurring during 1990–2010) in the CARDIA cohort was associated with worse midlife cognitive function. However, CARDIA lacked earlier-life cognition data, a major potential confounder. We replicate and extend this study in the National Longitudinal Survey of Youth (1979), which collected cognition measures in approximate adolescence. Following Grasset, we estimated associations between 20-year income volatility throughout early-to-mid adulthood (mean ages 27–47 in 1990–2010) and midlife cognitive function using confounder-adjusted linear regression models. Income volatility was assessed as the standard deviation (SD) of the percent change in income between successive annual or biennial surveys and as the number of income drops >25% between surveys. For cognition, we averaged z-scores on immediate and delayed 10-word recall, backwards counting, and serial 7's tasks. We found that income volatility and number of income drops were negatively associated with midlife cognitive function (e.g., each SD increase in income volatility was associated with 0.081 SD lower cognition z-scores, 95% CI: −0.121, −0.041) independent of earlier-life cognitive skills (e.g., before adjustment, $\hat{\beta }$ = -0.095, 95% CI: −0.136, −0.054). Our study adds to a growing literature suggesting life course financial volatility may adversely impact cognitive aging.