The objective of this paper is to estimate the causal effect of coresidence with an adult child on depressive symptoms among older widowed women in South Korea. Data from the first and second waves of the Korea Longitudinal Study of Aging were used. Analysis was restricted to widowed women aged ≥ 65 years with at least one living child (N=2,449). We use an instrumental variable approach that exploits the cultural setting where number of sons predicts the probability of an elderly woman's coresidence with an adult child but is not directly correlated with the mother's depressive symptoms. Our models adjust for age, education, total assets, residence, functional limitations, self-rated health, and various illnesses. Our robust estimation results indicate that, among older widowed women, coresidence with an adult child has a significant protective effect on depressive symptoms, but that this effect does not necessarily benefit those with clinically relevant depressive symptoms. Future demographic and social transitions in South Korea portend that older women’s increasing vulnerability to poor mental health is an important though less visible public health challenge.
Keywords: living arrangements, coresidence, depressive symptomatology, elderly, KLoSA