Koranic Schools in Senegal : A real barrier to formal education ? 1

State education systems in Sahelian countries do not teach religious education, which is provided by the informal sector. This article is a rst attempt to quantitatively study how this dual educational system works and whether the potential competition between both systems is a key factor behind low primary school enrolment in Senegal. The analysis is based on a unique national dataset covering 1,800 households, with detailed information on formal and Koranic schooling of 5 to 21 year-old children. In our sample, over half of the girls and 60% of the boys attend a Koranic school for at least one year, although most of them stay for only two to three years. We present a brief background on Islam and Koranic schools in Senegal to provide a better understanding of the complexity of the subject. We then examine the determinants of Koranic schooling before going on to analyze its compatibility with formal schooling. A descriptive analysis shows that children who attend Koranic school for a few years have a higher probability of attending formal primary school than those who do not go to Koranic school at all and those who pursue long Koranic studies. So as to identify the substitution between Koranic schools and formal school, we use an instrumental strategy based on the opening of formal schools. Our I.V. estimations show that substitution e ect dominates for boys. This substitution probably shows that both school careers are considered as relevant educational choices by some Senegalese households, and that there is competition between formal school enrolment and Koranic education in Senegal. As the opening of formal schools changes the formal school enrolment decisions, the choice of full time Koranic enrolment is probably partly due to the poor quality of formal schools. Hence, improving the quality of formal education could actually raise formal school enrolment. The existence of full time Koranic education could therefore be partly a signal indicating the poor quality of the formal school system. However it is probably the case that this existence is also partially based on real preferences for religious education. The preferences of some households for Koranic education could therefore prevent formal primary education even if the quality of formal education is good. The potential existence of such households would raise the question of an improved integrated education system facilitating them nding a balance between the two types of education.