Leaders play an important role in the performance of the organizations (Hermalin, 1998, 2013, Van der Heijden, Potters and Sefton, 2006). At the same time, leaders are quite mobile and often move to a different organization, retire or are dismissed. The replacement of leaders can bring advantages to an organization by rejuvenating it, bringing in new ideas or by disciplining leaders and decreasing the abuse of power (Datta and Rajagopalan 1998; Ocasio 1994). However, leader succession can also be damaging for the organization: the arrival of a leader who does not belong to the same group can decrease organizational performance. This paper focuses on the role of group membership of the leader on the performance of the organization. We ask: Do out-group leaders, exhibit a weaker motivation to work and are less productive than in-group leaders? Do group members anticipate this and cooperate less in groups led by an out-group leader compared with groups led by an in-group leader? Do these differences result in lower payout for group members with an out-group leader? Does the effect of group membership depend on the level of competence of the leader? To explore these questions, we used a modified public goods game in which the marginal per capita return from contributions to the public good (or MPCR) depends on the productivity of the leader in a real-effort task. We allowed the leader to be either part of the organization (in-group) or to be an outsider (out-group). In the first case, group members had interacted with the leader before. Out-group leaders, by contrast, do not share membership with members of the organization. They belong to a different group, so participants had had no direct or indirect interaction with them before. The second variation that we introduced relates to the leader-selection mechanism. We either selected a leader randomly (random leader) or selected a leader among the pool of the highest-performing participants in the session (high-performance leader). In the analysis, we consider the effect of those manipulations on leaders’ performance in a real-effort task, group members’ contributions to a public goods game and participants’ payoffs. We find that the group membership of the leader influences the performance of the organization. However, this depends on the leader-selection mechanism. When the leader is randomly selected, members of a group with an out-group leader received a lower payment than groups with an in-group leader.