Living Arrangements and Health of the Elderly in India

Abstract

Do multigenerational (co-residential) families have protective effects on elderly health? Demographic literature on aging in developing countries has started to examine this question as the contours of global population have been undergoing dramatic changes. Nevertheless, the theoretical and empirical literature on the relative benefits for the elderly of residing in multigenerational households versus living alone, have remained remarkably elusive. In part, the empirical inconsistency is a result of a significant methodological gap in the extant literature: most empirical studies are based on cross-sectional data where the authors have been largely unsuccessful in eliminating explanations based on the possible selection effects. India offers an interesting context to study this relationship as the country experiences a growing elderly population coupled with a severe lack of institutional systems of care for the aged. This dissertation draws data from the India Human Development Survey (2004-05) - a nationally representative, multi-topic data set of 41,554 households. It focuses on the relationship between household composition-whether the elderly are living independently, with children, or with other relatives-and short-term morbidity in the last month. The analysis uses standard multivariate regression models and a relatively unconventional technique-propensity score analysis to account for the endogeneity/selectivity problem. Three particularly salient conclusions are drawn from this research. First, household level analyses using propensity score methods highlight the importance of multigenerational families to the health of the elderly. These results also suggest health effects of household wealth, urban location, the number of adults in the household, and (male) gender. A second set of analyses show that multigenerational families also spend more on medical care when the elderly do get sick. Moreover the same set of household variables that predict better elderly health (wealth, urbanization, adults, gender) also predict higher medical expenditures. Finally, multilevel analyses, using district-level data from the Census of India (2001), corroborate the "urban advantage" finding and demonstrate that health of the elderly is affected not only by household compositional factors (e.g. living arrangements) but also by the larger context created by urbanization.