This paper analyses the relation between two important aspects of human capital: education and health. The contribution of our paper to the literature is three-fold: some further tests for causality in the relation between education and health are provided; it is tested whether results are affected by scale of reference bias and unobserved heterogeneity; and the results are used to calculate the health returns to education. We use data from a large survey for the Netherlands to estimate the education effects on health. Calculated at the average value of GDP per capita, the implied health returns to education are 1.3-5.8%. We further test for causality in the relation between education and health and for the effects of scale of reference bias and unobserved heterogeneity. None of these appear to affect our conclusions.