This study applies instrumental variable panel data techniques to estimate the effects of civil wars and conflicts on food security in developing countries. From a statistical standpoint, the results glaringly pinpoint the danger of using conventional panel data estimators when endogeneity is of conventional type, i.e. with respect to the idiosyncratic error term. However, from a policy perspective, we find that, in general, civil wars and conflicts are detrimental to food security, but the negative effects are more severe for countries unable to make available for their citizens the minimum dietary energy requirement under which a country is qualified for food aid.