Abstract: The ‘integrated hypothesis’ predicts that reconciliation (the post-conflict friendly interaction between former opponents observed in various group-living species) functions to reduce anxiety and the risk of aggression from the former opponent or a bystander in the aftermath of a conflict. It also predicts that relationship quality between opponents affects the occurrence of reconciliation and modulates the anxious response of the opponents after a conflict. Because of the asymmetric nature of aggressive interactions, the cost of aggression is likely to differ between the victim and the aggressor. The aim of this study was to test the predictions of the ‘integrated hypothesis’ independently for the victim and the aggressor of a conflict. We collected data on two wild groups of Barbary macaques. This study represents, to our knowledge, the first systematic test of the integrated hypothesis on wild, non-provisioned animals. Victims of aggression were at a greater risk of receiving aggression from the former opponent or a bystander after a conflict and showed elevated anxiety. We found no such costs for the aggressor. Reconciliation reduced anxiety in the victim but did not reduce their risk of receiving aggression. Finally, relationship quality affected the occurrence of reconciliation but did not modulate post-conflict anxiety. The results of our study show that the costs of aggression are asymmetrically distributed between the victim and the aggressor. Such differences are likely to lead to different social tactics used by the victim and the aggressor in the aftermath of a conflict.