Scholars debate whether climate change has a consistent effect on the likelihood of armed conflict in Africa. One major pathway by which climatic variability is hypothesized to increase conflict is by decreasing food availability. However, limitations on food access at both the local and national level in many developing African countries force most armed groups and communities to depend on locally produced food. These actors are therefore likely to use violence to establish control over more food resources or be stationed where more food is available, suggesting that food abundance might also be driving conflict. The present study employs novel data on wheat and maize yields in Africa measured at the very local level to empirically evaluate this hypothesis on a highly disaggregated conflict indicator. To account for the endogenous relationship between conflict and food production, average local levels of drought are used as an instrument. The findings show that, contrary to previous expectations, conflict is driven by higher yields, on average, and not by scarcity.

Food security has attracted widespread attention in recent years. Yet, despite preliminary evidence connecting food insecurity to political violence, we lack a systematic understanding of the relationship(s) between local food resources and violence against civilians. This study develops a food-security based theory to explain the significant variation that we observe in violence against civilians across both time and subnational geographic space. We argue that combatants, be they government or rebel actors, often must turn to local agricultural resources for sustenance. During times of relative peace, armed actors and civilians have long time horizons, and the prospects of repeated interactions thereby promote a strategy of co-optation to obtain food resources. However, the existence of immediate conflict in a region leads armed actors to discount the benefits of future interactions in favor of obtaining food immediately, using violence if necessary. In estimating a series of statistical models on a sample of all African countries (1997–2009), we find robust support for our expectations: cropland increases the frequency of violence against civilians during periods of conflict, but has an added pacifying effect during times of peace.

No abstract

No abstract