Threat Escalation and Crisis Stability: A Game-theoretic Analysis

Abstract

We develop a model of crisis stability based on a deterrence game. Players are assumed to be able to choose any level of initial cooperation or noncooperation; the more cooperative player (if there is one) may then choose to retaliate. In a crisis, players are assumed to have escalated their conflict and desire to stabilize it before it explodes. Players can, by threatening their opponent, stabilize the new status quo. In addition, points of threat escalation are identified at which neither player can prevent the opponent from escalating further without threatening the opponent more severely than before the crisis erupted, thereby heating up an already tense situation. Crisis stabilization is aided by being close to the full-cooperation position, though, paradoxically, both players may benefit from having created a crisis that only escalating threats may resolve. We discuss ways of avoiding threat escalation and the increased risk of war that it entails, especially in conflicts between the superpowers.