The paper studies the relationship between prevention and risk sharing in the optimal design of a climate policy. If countries could credibly commit to risk transfer mechanisms, heterogeneity in impacts would not be a problem. Instead, inefficient risk sharing looks like a second source of risk for the social planner. The question is then whether coun- tries should engage in more mitigation when they face heterogenous impacts. We determine conditions on risk aversion and prudence that guarantee an increase in mitigation under inefficient risk sharing. Generally speaking, mitigation tends to rise either when the event is catastrophic (i.e. low probability, high loss), or when the social planner is not very prudent and the inequality risk small. Contrary to intuition, risk sharing and prevention are usually complements. To induce a larger mitigation effort, countries should first of all solve the inefficiencies in the risk sharing mechanism.