The benefits of being accountable for task outcomes is well established. This is particularly true for leaders as being accountable can help counteract the disinhibiting effects of power. Despite this, it is unclear from past work whether leader accountability is beneficial, harmful, or even related at all to team member well-being. In this dissertation, I propose that whether leader accountability is beneficial or harmful depends on what part of their role leaders feel accountable for—their task or emotional impact. Specifically, I suggest that the established construct of leader accountability is better considered as leader task accountability—leaders feeling responsible to consider and justify how their reasoning and decisions impact the team’s task—and propose a novel construct of leader emotional accountability—leaders feeling responsible to consider and justify the impact of their affective expressions on team members. I argue that leader task accountability is negatively related to team member well-being whereas leader emotional accountability is positively related to team member well-being. To explain these effects, I draw on theory of organizational compassion and contend that leader compassion mediates these relationships. Lastly, I argue that leader task interdependence moderates these effects by amplifying them. I test my model across five studies: a scale validation pilot, two field studies, and two pre-registered experiments, and find general support. This dissertation contributes to research on accountability and leader emotions.