Most governance arrangements involve spatial units with highly unequal powers, for example, a feudal monarchy and its principalities, an empire and its colonies, a formal empire and an informal empire (or sphere of influence), a national government and its subnational entities, or a regional government and its local entities. In this situation, the dominant unit ( A ) usually enjoys some discretion about how to institutionalize its authority over the subordinate unit ( B ). An important element of this decision concerns how much authority should be delegated to the weaker unit. The authors simplify this dimension of governance along a continuum of “direct” and “indirect” styles of rule. Why, in some cases, does one find a relatively direct (centralized) system of rule and in others a relatively indirect (decentralized) system of rule? While many factors impinge on this decision, the authors argue that an important and highly persistent factor is the prior level of centralization existing within the subordinate unit. Greater centralization in B is likely to lead to a more indirect form of rule between A and B , all other things being equal. The authors refer to this as an institutional theory of direct/indirect rule. Empirical analyses of this hypothesis are applied to patterns of direct and indirect rule (1) during the age of imperialism and (2) across contemporary nation-states. The article concludes by discussing applications of the theory in a variety of additional settings.