This paper aims to assess whether a causal effect exists between maternal education and child survival in Madagascar. The omission of factors such as mother's health, innate ability and time preferences could lead to an overestimation of the true effect of education. The case of sub-Saharan Africa where child mortality rates are the highest, is overlooked by most of the causal evidence gathered so far for developing countries. The present paper attempts to redress this omission through the adoption of a careful empirical strategy. The analysis sheds light on the mechanisms at stake based on information on hygiene practices, housing conditions and the health care administered before, during and after childbirth. The results demonstrate that mothers' education has a positive and strong effect on their offsprings' survival probabilities. Wealth on its own has a strong effect but seems to account for only a third of the effect of maternal education. Abstract This paper aims to assess whether a causal effect exists between maternal education