This project aims to contribute to a better understanding of the causes of quality of government (QoG) -- here defined as an impartial and effective government -- in democracies, by focusing on the degree of decentralization as key independent variable. I argue that the current inconclusiveness in this literature stems, in large part, from employing incomplete indicators for QoG, not studying whether the impact of decentralization increases with duration, not including the tri-dimensional nature of decentralization, using mostly cross-sectional data and not considering the potential moderating roles of the levels of democracy and societal resources. In order to test thehypotheses thus derived, I will, in a first step, run a statistical analysis using a novel panel database, and, in a second step, perform a case study of countries in which the mechanisms uncovered are especially relevant, to understand the role of context and analyze other sub-dimensions of both variables. Supervisor: Tiago Fernandes (CEI-Iscte) Pedro Magalhães (ICS-UL)