The dissertation project is a cross-regional, cross-municipal study of local governance in Brazil. It tests the proposition that the autonomy of civic organization, the competitiveness of local political elites, and the interaction between these two variables structure differences in the quality of municipal citizen representation. Differing forms of political representation, in turn, shape the process and performance of local policy-making. I argue that while the autonomy of civil society is essential to the consolidation of democratic practices in polities dominated by patrimonial practices, civil society's capacity to do_ this is mediated by inter-elite competition. A dense, mobilized civil society gives entrepreneurial elites a support base and rationale to engage in reform. However, it is the additional factor of competitive local elite politics that leads politicians to defect from traditional political practices. Empirically, this study will gather data on civic associations, elections, and policy-making in three municipalities (Maceio, Recife, and Porto Alegre) located in two different regions of Brazil (the North-East and the South). It will cover a period from the mid-1980s through the 1998 municipal elections.