The paper studies the impact of ethnic diversity on social participation of young people We ¯rst propose a theoretical model in which the agents choose between structured and unstructured social activities given the ethnic composition of the groups they join. Our model predicts that segregation should have non negative e®ect on social participation, while fractionalization should have a non positive e®ect. Moreover, we argue that the latter results should be stronger where spontaneous participation is taken into account. We test our predictions using English census data together with the `Longitudinal Survey of Young People in England' (LSYPE) and we con¯rm the predictions of our model. Finally, we use an IV strategy based on both historical and geographical data to correct for endogenous sorting into neighborhoods. The results we get are even stronger than those obtained where the ethnic composition of the neighborhood is taken as exogenous.