Abstract: Pupils are selected into Malawi’s public secondary schools based on merit. Top performers on the primary school exam are selected into relatively well endowed conventional secondary schools (CSSs) while second tier performers are selected into community day secondary schools (CDSS). This paper exploits the merit based selection mechanism to estimate the impact of selection into CSS on pupil achievements. Regression discontinuity estimates show that pupils selected into CSS are less likely to drop out of school or repeat grades. This result shows that, in the Sub-Saharan context, appropriate school policy can encourage pupils to stay in school. Further results indicate that the higher probability of staying in school is to a large extent related to the higher ratio of well educated teachers in CSSs.