Because personalist dictators wreak havoc in their own countries, threaten neighbors, and set the stage for renewed dictatorship after they fall, the principal policy recommendation implied by our research is that international policy makers should avoid contributing to the personalization of dictatorial rule, even if security concerns suggest support. Dictators with unlimited policy discretion can switch sides easily and unpredictably, using the very weapons provided by their allies to turn against them. Decisions about economic and military intervention aimed at ending dangerous or abhorrent dictatorships should be informed by realistic assessments of whether the intervention is likely to succeed and what will happen if the dictator falls. After foreign intervention to oust a personalist dictator, the likelihood of democratization is not high. The more arbitrary, violent, and paranoid the personalist dictator, the more likely his overthrow will result in another autocracy, civil war, or a failed state. We suggest that personalist dictators who rely on narrow ethnic, clan, or religious groups for support are especially likely to experience bloody transitions and violent, unstable futures.