We study whether one reason behind female underrepresentation in leadership is that female leaders are less effective at coordinating action by followers. Two experiments using coordination games investigate whether female leaders are less successful than males in persuading followers to coordinate on efficient equilibria. Group performance hinges on higher-order beliefs about the leader's capacity to convince followers to pursue desired actions, making beliefs that women are less effective leaders potentially self-confirming. We find no evidence that such bias impacts actual leadership performance, identifying a precisely-estimated null effect. We show that this absence of an effect is surprising given experts' priors.