A realistic theory of economic sanctions should be built on the facts that sanctions are a game of issue linkage involving two or more issues, players may not know each other's preferences for the outcome of the game, and threatening sanctions may be as important as imposing sanctions as a strategy in international disputes. The threat and use of economic sanctions are modeled as a multistage game of two-sided incomplete information between a target and a coercer. The threat stage is critically important for understanding the outcome of sanctions, and current empirical studies suffer from a case selection bias. Economic sanctions are likely to be imposed when they are not likely to succeed in changing the target's behavior. Sanctions that are likely to succeed will do so at the mere threat of sanctions. Despite the unlikely success of sanctions, coercers must sometimes impose sanctions, even after the threat of sanctions has failed to change the target's behavior.