I use a dynamic microsimulation model to analyse the distributional effects of educational policies in Côte d’Ivoire. The simulations replicate several policies in force or subject to debate in this country. The direct effects between education and income as well as the different transmission channels, such as occupational choices and household composition, are analysed. The simulations show that even if the most optimistic policy considered was really set up the effects on poverty would be very modest as long as returns to education remain constant and the labour market segmented. This has interesting implications for the elaboration of poverty reduction strategies.