Recent empirical work scrutinizes the ability of economic sanctions to destabilize targeted leaders. Limitations in data and modeling choices, however, may have inflated estimates of sanctions’ efficacy. I propose a unified theoretical model, incorporating the possibility that leaders targeted with threats and imposed sanctions differ in baseline risks from those who are not. I combine this hazards approach with an empirical strategy to account for differences in ex ante risks and improved data on leader failure. This approach uncovers a considerably more modest effect. Sanctions rarely destabilize their targets.