Local warming and violent armed conflict in Africa

Research on the effect of climate change on violent armed conflict relies almost exclusively on analysing annual variation in climatic conditions. A shortcoming of this approach is that it conflates weather variation with climate change, while implicitly assuming that adverse weather shock could immediately trigger conflict. Although this relatively high-frequency data can help understand conflict seasonality, it fails to address the question of whether climate change is an important conflict determinant. This study tries to address this issue using long-term change in local climate to proxy climate change. Focusing on the African continent, shifts in average temperature and precipitation levels are used to estimate the effect on conflict risk between 2003–17. The data is analysed using Bayesian model averaging to test if the variables measuring local climatic conditions contribute consistently in explaining conflict risk. The reduced-form estimations show that temperature is robustly linked to armed conflict: a two-standard deviation increase in average temperature corresponds to about a 31 percent increase in conflict risk. Precipitation changes have no discernible effect. Changes in local climate are more strongly linked to the continuation of existing conflicts, rather than the outbreak of new ones. The association between climate and conflict found in the analysis also suggests a potential lack of adaptation. While the findings of this study are in agreement with earlier results, one remaining shortcoming is that the analysis does not provide much insight into the specific mechanisms linking climate and conflict.