There is a strong, positive and well-documented correlation between education and health outcomes. There is much less evidence on the extent to which this correlation reflects the causal effect of education on health - the parameter of interest for policy. In this paper we attempt to overcome...