📌 Main Argument

Surrender is explained as a contagious collective-action problem: victory in battle depends on soldiers choosing to fight together rather than flee, but each soldier’s willingness to fight hinges on expectations about whether comrades will do the same. Soldiers look to what others did in comparable situations, so mass surrender in one engagement raises the probability of surrender in later battles. When no recent precedent exists, widespread capitulation is unlikely.

🧭 Evidence: New Data on Conventional Battles, 1939–2011

• A new dataset covering conventional battles in all interstate wars from 1939 to 2011 is used to test the theory.
• Empirical analysis of this dataset provides support for the claim that battlefield surrender spreads via precedent: prior mass capitulations predict higher likelihoods of surrender in subsequent battles when combatants face similar conditions.

🔍 Key Findings

• Surrender behaves like a contagious social phenomenon across battles rather than an entirely isolated decision.
• Individual decisions to fight or flee are shaped by expectations formed from recent battlefield precedents.
• The absence of recent mass-surrender precedents makes coordinated flight less likely and prolonged fighting more likely.

⚖️ Why It Matters

These results deepen understanding of battlefield resolve and reveal how micro-level expectations aggregate into life-or-death outcomes. The findings have broader implications for designing political-military institutions and for strategic decisions to initiate, continue, or terminate wars, since the likelihood of capitulation depends in part on observable precedents rather than only on material capabilities or immediate battlefield conditions.