DP19990 Education, Lifestyles, and Health Inequality
We study the effect of lifestyles on the education gradient of life expectancy. We use panel data on health behavior and health outcomes to estimate latent lifestyle types and their impact on health dynamics. We find that the higher frequency of health-protective lifestyles among the more educated individuals explains almost 1/2 of the education gradient in life expectancy. To understand lifestyle formation, we build a life cycle model where lifestyles and education are jointly chosen early in life. These two investments are complementary because of the more educated’s higher income and the higher yield of their health-protective behavior. Importantly, with these complementarities, individuals with lower costs of healthier lifestyles self-select into higher education. Quantitatively, we find the three mechanisms similarly important in explaining the correlation between education and healthy lifestyles. We also find that the increase in the college wage premium over the last decades has widened the education gradient in lifestyles, resulting in a one-year increase in the education gradient of life expectancy across cohorts born in the 1930s and 1970s. Of this increase, 40% is driven by the direct effect of wage changes and 60% by the induced changes in the composition of college graduates and high school dropouts.
