This dissertation examines the effect of fiscal decentralization on poverty reduction and explores potential transmission channels through pro-poor sectoral outcomes such as basic education, basic healthcare and agricultural productivity. We first develop a theoretical model to explain the interaction between decentralization and poverty reduction outcomes. In particular, we show that the marginal effect of fiscal decentralization on pro-poor sectors depends largely on the outcome of the trade-off between potential benefits derived from better matching of local preference due to local proximity, and the lack of technical capacity at the local level. This finding provides, in a way, a theoretical explanation of the different outcomes observed in fiscal decentralization programs around the world. This inconclusive theoretical result motivates an empirical analysis to assess whether there is any statistical significant relationship between fiscal decentralization and poverty. We implement this estimation using the Generalized Method of Moment Instrumental Variable (GMM-IV) methodology on 97 countries spanned over the period 1975-2000. Our estimation results reveal a statistically significant, but non-linear relationship between fiscal decentralization and poverty. In addition, we empirically explore potential transmission mechanism of the effect of fiscal decentralization on poverty through three sectors (basic education, basic healthcare and agricultural extension) that have been shown in the literature on basic needs and also by development practitioners to have significant bearing on the well-being of the poor. Finally, given the relatively high level of poverty in addition to the fact that most African countries are far behind in attaining their Millennium Development Goals (MDGs) targets, we investigate whether the effect of fiscal decentralization on poverty will be greater in Sub-Saharan Africa (SSA) relative to other regions.