Humanitarian aid has long been considered an important means to reduce hunger and suffering in developing countries. A recent finding by Nunn and Qian (US food aid and civil conflict, American Economic Review 2014; 104: 1630-1666) that such aid from the US increases the incidence and duration of civil conflict in recipient countries, however, questions the effectiveness of this policy and poses a serious policy concern for the US government. We revisit this issue by conducting a successful replication study of the results in their paper. In order to further scrutinize their claims that a heterogeneous effect of food aid on conflict is not present, we employ a semiparametric endogenous estimation procedure. We show that their parametric models cannot be rejected and argue that their findings are robust.