The impact of health insurance on residents' health is one of the focal points of academic research. Due to the fact that China's medical insurance system is composed of a variety of programs and that the pooling districts are at the lower administrative level, enrollment in different medical insurance programs or at different places may have certain influences on the health of residents. This has mostly been neglected by previous studies. This paper uses data from the 2015 China Migrants Dynamic Survey (CMDS), focusing on the senior floating population and taking the difference in government subsidy proportions as an instrumental variable in order to identify the effects of health insurance programs and regional differences on the health of the senior floating population. Three effects were observed: First, participation in the health insurance system significantly improves floating seniors' self-rated health. Second, the health status of floating seniors affects their choice of health insurance program: Less healthy persons tend to choose high-paying, wide-coverage basic medical insurance available for urban employees. Using an instrumental variable to control for the problem of endogeneity, it is discovered that compared with the basic medical insurance system for urban residents, the system for urban employees significantly enhances the health of the senior floating population. Third, "adverse selection" could be observed in the choice between enrolling in health insurance at the place of settlement or another place. Senior migrants with worse self-rated health tend to choose place of settlement in order to enjoy higher compensation and less complex reimbursement procedures. With an instrumental variable to control for the problem of endogeneity, it was found that compared with joining the medical insurance system at other places, joining at a place of settlement could improve the health of the floating senior population.
Keywords: China; adverse selection; floating population; health; medical insurance system.