This paper explores why some sectarian providers distribute welfare goods broadly, even to non-coreligionists, while others allocate services more narrowly. Focusing on Lebanon, the study compares the welfare distribution strategies of Christian, Shi’i Muslim, Sunni Muslim and Druze political parties. In Lebanon, a quintessential case of a weak state where power-sharing arrangements ensure the political salience of religion and ethnicity, welfare is a terrain of political contestation, but the distribution of welfare goods varies across parties. Many assume that sectarian providers merely “take care of their own,” ensuring the well-being of co-religionists to the neglect of others. Alternatively, some contend that sectarian parties do not exert great efforts to serve their communities because they are virtually assured the support of co-religionists in a sectarian political system. Contrary to these approaches, I argue that two key factors shape how service provision is used to build political support and, specifically, who is brought into the party or movement: first, whether a party has achieved dominance within its respective sect; and second, whether it engages in electoral, mobilizational or militia competition. Once a sectarian party has achieved dominance within its sect and when it contests elections, it plays a “double game,” shoring up a core, largely co-religionist group of supporters with broad welfare benefits while extending some types of services to marginal supporters. Conversely, when a party faces competition from other parties within its sect and when it aims to mobilize mass support or engages in militia fighting, it tends to focus services on a hard core of co-religionist supporters who may be members of militia fighter families. The research for this paper is based on data collected during multiple field research trips to Lebanon in 2004 and between 2006 and 2008 and employs multiple methods of data collection and analysis, including Geographic Information Systems (GIS) analysis, an original national survey of Lebanese citizens on access to social welfare (n=2,000), standardized open-ended interviews with representatives from political parties, religious charities, NGOs, government ministries as well as local academics, development consultants and journalists (n=157), standardized open-ended interviews with beneficiaries of social welfare programs conducted by a team of trained and well-qualified Lebanese graduate students matched with co-religionist respondents (n=135), and archival research.