When providing social accounts (Sitkin & Bies, 1993) for the unethical conduct of subordinates, leaders may use language consistent with cognitive strategies described by Bandura (1991, 1999) in his work on moral disengagement. That is, leader’s social accounts may reframe or reconstrue subordinates’ unethical conduct such that it appears less reprehensible. We predict observers will respond negatively to leaders when they use moral disengagement language within social accounts and, specifically, observers will ostracize these leaders. In addition, we predict that observer moral disengagement propensity moderates this effect, such that the relationship between leaders’ use of moral disengagement language within a social account and ostracism is stronger when observer moral disengagement propensity is lower versus higher. Finally, we predict that the reason why observers ostracize the leader is because observers perceive the leader’s social account with moral disengagement language as unethical. Thus, perceived leader social account ethicality is predicted to mediate the interaction effect of leader’s use of moral disengagement language within social accounts and observer moral disengagement propensity on ostracism. Results from an experiment and field study support our predictions. Implications for theory and practice are discussed.