We provide new evidence on the causal effect of education on health. Our empirical strategy exploits the reduction in access to college experienced by individuals reaching college age shortly after the 1973 military coup in Chile, which led to a sharp downward kink in enrollment for the affected cohorts. Using data from the vital statistics for the period 1994-2017, we document an upward kink in the age-adjusted yearly mortality rate for these cohorts, a pattern that we also observe in matched individual-level records for hospitalized patients. Leveraging the downward kink in college enrollment, we estimate a negative effect of college on mortality, which is larger for men, but also sizable for women. Affected individuals have worse labor market outcomes, lower income, and are more likely to be enrolled in the public health system. They also report lower consumption of health services, which suggests that economic disadvantage and limited access to care plausibly contribute to the effect of education on health.