Abstract
Health influences many economic outcomes, but its dynamics are still poorly understood. We used k-means clustering, machine learning techniques, and data from the Health and Retirement Survey to identify health types in midlife and old age. We identified five health types: vital resilient, fair health resilient, fair health fragile, frail resilient, and frail fragile, which are characterized by different initial health states and health and mortality trajectories. Our five health types account for 84% of the variation in health trajectories and are explained by past health dynamics rather than observable characteristics such as age, marital status, education, sex, race, health-related behaviors, or health insurance status. We also show that health types are an important driver of health and mortality heterogeneity and dynamics. Our results highlight the importance of better understanding the formation of health types and modeling it appropriately to properly assess the impact of health on people's decisions and the implications of policy reforms.
