What Levels the Association Between Income and Mortality in Later Life: Age or Health Decline?

Researchers frequently use the "age-as-leveler" hypothesis to explain decreasing inequality and a weakened relationship between socioeconomic position and health in old age. This study examined whether health status can explain the age pattern in the association between income and mortality as predicted by the age-as-leveler hypothesis. This study used longitudinal (1991-2002) data from the SWEOLD and LNU surveys. The analytical sample consisted of 2,619 people aged 54-92 in 2003. Mortality (2003-2014) and income (1991-2000) was collected from Swedish national registers. Poisson regression was used to estimate associations between mortality, income, age, and health status. Average marginal effects were used to visualize interaction effects between income and age. The association between income and mortality weakened in those aged 84 and older. However, health status explained a large part of the effect that age had on the association between income and mortality. Analyses done after stratifying the sample by health status showed that the association between income and mortality was strong in people who reported good health and weak or nonexistent in those who reported poor health. Age leveled the income-mortality association; however, health status, not age, explained most of the leveling.