Crisis Bargaining and Escalation to War

Core Idea

Bargaining models explain war as resulting from private information asymmetries, commitment problems, or indivisibilities. Crises escalate when states use threats to communicate but face incentives to bluff, creating risks of inadvertent war. War can occur even when all parties would prefer peaceful settlement ex ante.

Explainer

The bargaining model of war begins from a puzzle that your prerequisite security dilemma analysis does not fully resolve: war is mutually destructive. Both sides pay in lives and resources, and the outcome is uncertain. Given this, there should almost always exist some negotiated settlement that both sides would prefer to fighting — a deal that gives each side what they could expect to win from war, minus the costs and risks of fighting. So why does war ever occur? The bargaining model offers a rigorous answer using the probability and expected-value tools in your prerequisites: war happens because the conditions required for a negotiated settlement fail in specific, identifiable ways.

The first failure mode is private information with incentives to misrepresent. Each state knows things about its own military capability, resolve, and domestic political constraints that the other cannot directly observe. And crucially, each has incentive to *misrepresent* these things — to bluff about being stronger or more committed than it actually is. If both sides bluff and neither calls the other's bluff credibly, each may calculate that it would win the war, leading both to reject settlements that a fully informed observer would accept. The result is a war neither informed party would have chosen. This is why costly signals matter: burning bridges, mobilizing troops at great economic cost, making public threats that damage a leader's reputation if unfulfilled — these actions are credible precisely because they are costly to take. A cheap signal (just saying "I'm committed") provides no information; a costly signal moves beliefs.

The second failure mode is commitment problems. Even when both sides could identify a mutually preferred settlement, one side may be unable to credibly commit to honoring it in the future. The canonical case is a rising power: a state growing rapidly stronger relative to its rival. A settlement that satisfies both parties today will be renegotiated by the rising power tomorrow when it is stronger. The declining power knows this; rather than accept an arrangement it cannot enforce long-term, it may prefer to fight now while it still can. This logic explains why power transitions are historically dangerous — not because states are inherently aggressive, but because credible long-run commitments become structurally impossible when power is shifting rapidly.

The third failure mode is issue indivisibility. Some disputes cannot be resolved by splitting the difference. Control of Jerusalem, the Falkland Islands, or succession to a throne cannot be divided 60/40 — either one side controls it or the other does. When both sides value the indivisible prize highly and neither can credibly commit to sharing, war may be the only available mechanism. Crises are the high-stakes contexts where all three failure modes interact under time pressure: states issue threats and ultimata, each trying to signal resolve while managing the risk that the other side calls the bluff. Brinkmanship — deliberately creating the risk of catastrophic war to coerce a concession — is the extreme expression of this logic, and it was the dominant strategic posture of the Cold War. The policy implication is direct: reducing war requires addressing information asymmetries (transparency, verification, arms control), building commitment devices (international institutions, reputational investment), and finding ways to make apparently indivisible issues divisible through creative institutional arrangements.