The evidence of coming climate change has generated catastrophe-like statements of a future where a warmer, wetter, and wilder climate leads to a surge in migrant streams and gives rise to new wars. Although highly popular in policy circles, few of these claims are based on systematic evidence. Using a most-likely case design on Kenya 1989–2004, with new geographically disaggregated data on armed conflicts below the common civil conflict level, this study finds that climatic factors do influence the risk of conflicts and violent events. The effect is opposite to what should be expected from much of the international relations literature; rather, it supports the observations made by recent anthropological studies. Years with below average rainfall tend to have a peaceful effect on the following year and less robustly so for the current year as well. Little support is found for the notion that scarcity of farmland fuels violence in itself or in election years. More densely populated areas – not areas with a low land per capita ratio – run a higher risk of conflict. Election years systematically see more violence, however. The findings therefore support the notion that large-scale intergroup violence is driven by calculation and political gain rather than desperate scrambles for scarce land, pasture, and water resources.