We provide new evidence on the causal effect of higher education on mortality. Our empirical strategy exploits the reduction in college openings introduced by the Pinochet regime after the 1973 coup in Chile, which led to a sharp downward kink in college enrollment among those cohorts reaching college age in the following years. Using administrative data from the vital statistics, we document an upward kink in the age-specific yearly mortality rate of individuals in the affected cohorts. We estimate a negative effect of college on mortality between ages 34-74, which is larger for men, but also sizable for women. Individuals in the affected cohorts experience worse labor market outcomes, are more likely to be enrolled in the public health system, and report lower consumption of health services. This suggests that economic disadvantage and limited access to care play an important mediating role in the link between higher education and mortality.