This dissertation explains particularistic spending in the provision of local public goods by subnational governments in Mexico and Argentina. The main argument is that particularistic spending is driven by the strength and extension of territorial control by governors' political allies and rivals in local governments. Governors provide regions with particularistic benefits in local public goods that differ in targetability depending on whether they are governed by allied or rival mayors because targetable and non-targetable local public goods are not equally susceptible to successful credit claiming attempts by local politicians. First, credit for local public goods that benefit clearly geographically circumscribed communities can be claimed by local mayors, while credit for local public goods that benefit more than a single municipality cannot be credibly claimed by local authorities. Second, depending on their party affiliation, mayors are governors' political allies or rivals; and third, according to the size of the population they govern, mayors may mobilize strong or weak party organizations at election time. This study challenges models of particularistic spending centered on voters' characteristics and shows that governors provide particularistic benefits in targetable services to strong allies and that they predominantly favor weak allies and rivals when delivering non-targetable services. This dissertation bridges levels of analysis by exploring the effects of province-wide political competition on particularistic spending across regions. It shows that stronger territorial control by opposition mayors intensifies particularistic strategies. As a governors' rivals gain ground, her stronger allies obtain more particularistic benefits in targetable goods, and her weaker allies and rivals benefit more in non-targetable local public goods. Competition has a curvilinear effect on particularistic spending on targetable services, and a positive linear effect on non-targetable services. This work relies on three main sources of data: an original dataset on spending on seven areas of local public goods between the early 1990s and 2005 in four states in Mexico and four provinces in Argentina; an original survey of subnational legislators; and 150 in-depth interviews with subnational politicians. The study tests hypotheses through statistical analyses and provides qualitative evidence of particularistic spending on local public goods by specific administrations.