Can imposers of sanctions end economic coercion without the fear of strengthening their targets’ capabilities? Senders may prefer to end sanctions given its ex post inefficiency, yet doing so might provide the target greater access to resources and contribute to its offensive behavior. Targets’ inability to credibly commit to reversing their policies while enjoying the gains from sanctions relief, coupled with the difficulty of perfectly observing their compliance behavior, creates an obstacle for ending sanctions and resuming profitable economic transactions. Using a game-theoretic model of sanctions removal under uncertainty about targets’ intentions and compliance behavior, I formally demonstrate and empirically find that sanctions are more likely to end if senders can successfully detect targets’ compliance, but only if the target considers the promised sanctions relief attractive. Targets that offset the costs of sanctions will not value the promised sanctions relief and choose not to negotiate over sanctions removal.