Is Temperature Exogenous? The Impact of Civil Conflict on the Instrumental Climate Record in Sub‐Saharan Africa

Research into the effects of climate on political and economic outcomes assumes that short‐term variation in weather is exogenous to the phenomena being studied. However, weather data are derived from stations operated by national governments, whose political capacity and stability affect the quality and continuity of coverage. We show that civil conflict risk in sub‐Saharan Africa is negatively correlated with the number and density of weather stations contributing to a country's temperature record. This effect is both cross‐sectional—countries with higher average conflict risk tend to have poorer coverage—and cross‐temporal—civil conflict leads to loss of weather stations. Poor coverage induces a small downward bias in one widely used temperature data set, due to its interpolation method, and increases measurement error, potentially attenuating estimates of the temperature–conflict relationship. Combining multiple observational data sets to reduce measurement error almost doubles the estimated effect of temperature anomalies on civil conflict risk.