Subnational regimes vary significantly within countries around the world. Even in well-established national democracies " like the USA in the 1950s or Argentina in the 1990s " local regimes are often far from democratic. In the last five decades scholarship on political regimes has produced an impressive body of research, but the focus has been mainly national. This dissertation tackles what Robert Dahl called the "grave omission" of subnational regimes by providing and justifying rigorous descriptive and causal inferences for the 24 provinces of Argentina between its re-democratization in 1983 and 2007. At the level of description, I provide a detailed operationalization of the concept of subnational democracy, including a "thick" conceptualization and two alternative measurement strategies. The first (objective) strategy resorts to electoral and institutional indicators to create an index for all provinces and all gubernatorial terms since 1983. The second (subjective) strategy is based on the Survey of Experts on Provincial Politics (SEPP), in which 155 experts in all the provinces provided disaggregate information about their regimes for the period 2003-2007. Seventeen indices of different aspects of democracy were derived from this dataset. At the level of explanation, I propose a rentier theory of subnational democracy, which shifts the focus of the rentier-state literature up by climbing Sartori's ladder of abstraction from the concept of resource rents to that of fiscal rents. Drawing on fiscal theories of the state, I argue that inter-provincial regime differences are to a large extent explained by a type of rentierism that is not geographically determined by natural resources but politically created by certain fiscal federalism arrangements. I posit that less democratic regimes are more likely in rentier provinces " those that receive disproportionately large central government transfers and practically forgo local taxation. Intergovernmental revenue-sharing rules that produce large vertical fiscal imbalances and favor the economically smaller districts provide their incumbents with generous fiscal federalism rents that allow them to restrict democratic contestation and weaken checks and balances. Statistical evidence from the Argentine provinces supports this explanation, which overshadows its main alternative, modernization theory. These findings are robust to alternative estimators and measurement choices. Qualitative and quantitative evidence suggest that the effects of heavy public spending and statism on the economic autonomy of political actors are the main causal mechanisms at work.