Abstract
Population ageing poses growing challenges to social protection systems, particularly in contexts characterized by informal labour trajectories and uneven pension coverage. This study examines the association between current employment and multidimensional well-being among older adults of retirement age in Ecuador, focusing on heterogeneity across well-being domains and the relative importance of economic sufficiency and institutional access. A quantitative, non-experimental, cross-sectional study was conducted with a non-probability sample of 365 adults aged 65 years or older across the seven parishes of Rumiñahui canton. Multidimensional well-being was constructed as an equally weighted composite of four standardized domains: Health, Urban Environment and Accessibility, Relational Well-Being, and Household Conditions. Ordinary least squares models with HC3 robust standard errors were estimated, adjusting for sociodemographic characteristics, log-transformed monthly income, and access to health care and public transportation. Employment showed heterogeneous domain-specific associations: it was positively associated with Health (β = 0.359, Holm-adjusted p < 0.001) and Relational Well-Being (β = 0.155, Holm-adjusted p = 0.008), negatively associated with Urban Environment and Accessibility (β = −0.293, Holm-adjusted p < 0.001), and not significantly associated with Household Conditions after multiple-testing correction (β = 0.085, Holm-adjusted p = 0.067). In the fully adjusted global model, employment remained positively associated with Global Well-Being (β = 0.124, p = 0.014), although income showed the strongest standardized association (β = 0.397, p < 0.001), followed by health-care access (β = 0.254, p < 0.001); public-transport access was not independently associated with the global outcome. These findings indicate that employment at retirement age is associated with later-life well-being in a domain-specific rather than uniform manner and that economic sufficiency and health-care access are particularly salient correlates of overall well-being. Given the cross-sectional design, the findings should be interpreted as associations rather than causal effects.