The causal effects of education on health over the life course: evidence from Canada.

Objectives: The present article assesses the causal effects of education on health in Canada by gender.

Study Design: This is an empirical study of the relationship between education and self-rated health. Probability samples of cohorts born between 1935 and 1965 are observed in 1990 and 2011. The average age of the cohorts is 37 years in 1990 and increases to 58 in 2011.

Methods: Education is instrumented using the changes in school-leaving age laws. Both linear and two-stage least squares regressions are estimated.

Results: A positive health effect for education is detected, which is larger for women during the first half of life. For the second half of life as well, a positive causal effect for education is found for both men and women. But no strong evidence of gender difference in education-health gradient is found later in life.

Conclusions: The results highlight the lingering importance of education for health policy, as longer schooling seems to generate non-monetary benefits in terms of health externalities across the life course.