Working Paper: CEPR ID: DP15542
Authors: Ulrich J. Eberle; Dominic Rohner; Mathias Thoenig
Abstract: This paper investigates the impact of climate shocks on violence between herders and farmers by using geolocalized data on conflict events for all African countries over the 1997-2014 period. We find that a +1°C increase in temperature leads to a +54% increase in conflict probability in mixed areas populated by both farmers and herders, compared to +17% increase in non-mixed areas. This result is robust to controlling for the interaction between temperature and ethnic polarization, alternative estimation techniques, disaggregation levels, and coding options of the climatic/conflict/ethnic variables. We then quantify the impact on conflicts of projected climate change in 2040. We find that, in absence of mixed areas, global warming would increase total annual conflicts by about a quarter in whole Africa; when factoring in the magnifying effect of mixed settlements, total annual conflicts are predicted to rise by as much as a third. We also provide two pieces of evidence that resource competition is a major driver of farmer-herder violence. Firstly, conflicts are much more prevalent at the fringe between rangeland and farmland --a geographic buffer of mixed usage that is suitable for both cattle herding and farming but is particularly vulnerable to climate shocks. Secondly, information on groups' mobility reveals that temperature spikes in the ethnic homeland of a nomadic group tend to diffuse its fighting operations outside its homeland, with a magnified spatial spread in the case of conflicts over resources. Finally, we show that violence is substantially reduced in the presence of policies that empower local communities, foster participatory democracy, enforce property rights and regulate land dispute resolution.
Keywords: violence; farmer-herder conflict; climate change; temperature; nomadic; resource competition; Africa; Sahel
JEL Codes: D74; N47; O13; Q34; Z13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| 1°C increase in temperature (Q54) | 54% increase in conflict probability in mixed areas (Q34) |
| 1°C increase in temperature (Q54) | 17% increase in conflict probability in non-mixed areas (Q34) |
| mixed settlements presence (R23) | increase in conflict probability (D74) |
| absence of mixed settlements (R28) | 26% increase in climate-induced conflicts (Q34) |
| including mixed settlements effect (C21) | raises climate-induced conflicts estimate to 33% (Q34) |
| resource competition (Q34) | significant driver of violence (J12) |
| supportive policies (J08) | decrease in violence (D74) |