There is a moral presumption against the possession or exercise of coercive power, but political power is coercive. If we think that political power is sometimes morally justified, then given the moral presumption against coercive power, we need a theory as to when political power is morally justified and what justifies it. That is, we need a theory of political legitimacy. This dissertation develops and defends a novel theory of legitimacy, the Act-Focused Consequentialist Theory of Legitimacy. This theory departs in significant respects from existing theories. In Part One of the dissertation, “Four Theses on Legitimacy,” I argue that these departures are well-motivated given the shortcomings of these existing theories. I object that these theories are mistaken to attempt to ground the authority of the State or a moral obligation to obey the law because it is the law. Such attempts are mistaken because authority and legitimacy ought to be addressed separately. I also object that existing theories of legitimacy are radically incomplete insofar as they focus almost exclusively on the legitimacy of ‘the State’ at the expense of the legitimacy of particular actions by state actors. Furthermore, I articulate significant, enduring problems for theories of legitimacy that claim it is the ‘will of the people’ or the ‘consent of the governed’ (in some sense) that legitimates political power. In light of these objections, I argue in defense of a new approach to theorizing about legitimacy, an approach that ignores questions of authority, that primarily focuses on the legitimacy of particular actions rather than the legitimacy of the State, and that eschews grounding legitimacy in ‘the will of citizens’ in some sense. In Part Two, “An Act-Focused Consequentialist Theory of Legitimacy,” I develop and defend a theory that embodies this novel approach. I argue that what legitimates particular actions—and, ultimately, the State—is the production of good consequences. I argue that the good consequence to be produced is Functional Autonomy, the ability to successfully act on one’s authentic values, at least up to some reasonable threshold. To maximize overall Functional Autonomy, I argue that there ought to be two divisions of labor. First, there ought to be an institutional division of labor, according to which different institutions perform tasks integral to the production of Functional Autonomy. Second, there ought to be a role-based division of labor, a division of labor within those institutions according to which different state actors deliberate and act in specialized ways. Deliberating and acting in specialized ways amounts to deliberating and acting on role-based reasons. On this theory, a particular action by a state actor is legitimate if and only if the role-based reasons to perform the action are not outweighed by role-independent moral reasons. I propose that there is very strong moral reason to maximize Functional Autonomy, and so state actors will be morally justified in exercising coercive power so long as their actions are appropriately contributing to their institutional tasks, which in turn are producing Functional Autonomy. As for macro legitimacy, states enjoy degrees of legitimacy, where the degree of legitimacy is determined by how much Functional Autonomy a state produces. A recurring theme in this second part of the dissertation is that various problems that arise for consequentialism as a theory of interpersonal morality do not arise in the context of an act-focused theory of political legitimacy.