Less schooling is associated with increased biological risks for chronic disease, but whether increasing years of schooling through policy interventions reduces these risks remains unclear. We examine the effect of a major education reform introduced in 1959 that raised the minimum school leaving age from 14 to 16 years in France, offering a unique natural experiment. We assess the causal impact of increased schooling duration on 16 biomarkers of cardiovascular, metabolic, organ and immune function in a large cohort of men and women born around 1953. Using a Regression Discontinuity Design, we find that the reform led to a significant increase in schooling duration among children from disadvantaged families; but longer schooling did not translate into better biomarker profiles in adulthood. Eligibility to the reform had no impact on the biomarker profile of respondents from intermediate or high social class families, while it led to increased blood pressure and white cells counts in adulthood among those from low parental social class. These findings were robust across several sensitivity analyses. They emphasize the importance of considering the institutional context and the respondents' social origins when evaluating the health effects of compulsory schooling reforms. Our results do not necessarily question the premise that education leads to better health, but they suggest that law-mandated increases in schooling alone may not improve the health of disadvantaged groups.