Humans behavior often varies depending on the opponent’s group membership, with both positive consequences (e.g., cooperation or mutual help) and negative ones (e.g., stereotyping, oppression, or even genocide). An influential model developed by Hammond and Axelrod (HA) highlighted the emergence of macrolevel “ethnocentric cooperation” from the aggregation of microlevel interactions based on arbitrary tags signaling group membership. We extended this model to include a wider set of agents’ behaviors including the possibility of harming others. This allowed to check whether and under which conditions xenophobia can emerge beside or in alternative to ethnocentric cooperation. The model was compared to Swedish data documenting social unrest and proxies of cooperative behaviors at the municipal level. The validation results supported the model predictions on conflict but not the ones on cooperation, casting doubts on HA’s original argument.