Income serves as an indicator of success relative to others, and provides individuals with resources that strengthen their capability to face challenges and benefit from opportunities. Out of all social determinants of health, income is one of the strongest predictors of health outcomes. The positive association between income and health in the working-age population is well established; those with higher income tend to have better health. Less is known about the association between income and health among older persons. Several studies have observed that in old age, health inequalities decrease and the relationship between income and health weakens. However, at what point in the ageing process the association starts to weaken, and to what extent, is debated.

The ageing process highlights the need for several theoretical considerations in studies on income and health. Societies are stratified by age, as manifested through transitions in and out of education, work, and retirement. Moreover, the individual experience of the ageing process involves biological processes of decline. Many health problems, and particularly death, are uncommon events during most of adulthood. In old age, however, health decline becomes a normative experience, and in Sweden, more than 90 percent of all deaths occur among people aged 65 or older. The characteristics and magnitude of age-related changes in the association between income and mortality constitute one of the prime concerns in this thesis. I have used two contending perspectives to understand the empirical results: the cumulative (dis)advantage theory and the age-as-leveler hypothesis.

In this thesis, I have investigated the association between income and mortality across ages, with a focus on later life. More specifically, I studied the shape and magnitude of the income-mortality association across the life course. Furthermore, I tested two potential mechanisms that may shape this association in old age: health decline and mortality selection. Overall, this thesis shows to what extent and in what ways the association between income and mortality is maintained in old age

This thesis includes four empirical studies. Study I, II, and IV are based on data from Swedish national registers (n = 801,017 – 5,011,414). Study III used survey data (LNU and SWEOLD) linked with data from administrative registries (n = 2619). The results from Study I showed that the income-mortality association was curvilinear with diminishing returns of income in both mid-life and late-life. Study II showed that relative mortality inequalities in income decreased with age and absolute mortality inequalities in income increased with age up to age 85-90, after which the mortality inequalities decreased. The results from Study III showed that health decline partly explained the weakened income-mortality association among the oldest old. Finally, Study IV showed that selective mortality had a substantial impact on poverty rates in old age; poverty rates were consistently and substantially lower for those who survived than those who died. Furthermore, selective mortality had the largest influence on the surviving population when mortality rates were at their highest.