We study whether community based interventions improve risk sharing in rural Malawi. We focus on improvements in risk sharing arising as a result of design features of these interventions. We use a 2 year panel dataset collected within a cluster randomised trial to investigate whether two community based interventions - a participatory women’s group intervention and a home-based infant feeding intervention - targeted at improving reproductive and infant health, improve risk sharing amongst households following crop losses. We find strong evidence that such interventions help protect household consumption against a crop loss. Investigation of underlying channels rules out direct health effects and income smoothing, and suggests that increased social interactions due to the interventions may be the main driver.