Surprisingly high levels of within-group cooperation are observed in conflict situations. Ex- periments confirm that external threats lead to higher cooperation. The psychological literature suggests proximate explanations in the form of group processes, but does not explain how these processes can evolve and persist. We provide an ultimate explanation, in which cooperation is a rational response to an external threat. We introduce a model in which groups vary in their willing- ness to help each other against external attackers. Attackers infer cooperativeness of groups from members’ behaviour under attack, and may be deterred by a group that bands together against an initial attack. Then, even self-interested individuals may defend each other when threatened in order to deter future attacks. We argue that a group’s reputation is a public good with a natural weakest-link structure. We extend the model to cooperative and altruistic behaviour in general.
