In this paper we provide new evidence on the causal effect of education on adult depression and cognition. Using SHARE data, we use schooling reforms in several European countries as instruments for educational attainment. We find that an extra year of education has a large and significant protective effect on mental health: the probability of suffering depression decreases by 6.5 percent. We find a large and significant protective effect on cognition as measured by word recall. We also explore whether heterogeneity and selection play a part in the large discrepancy between OLS and IV (LATE) estimates of the effect of education on depression and cognition. Using the data available in SHARELIFE on early life conditions of the respondents such as the individuals’ socioeconomic status, health, and performance at school, we identify subgroups particularly affected by the reforms and with high marginal health returns to education.