How do citizens vote in rural Pakistan, and how much agency do they have in relation to local landlords, patrons and kinship networks in making electoral decisions? I explore this question in this dissertation through an empirical investigation of the voting behaviour of Pakistan's rural majority in its most populous and politically important province, Punjab, using a mix of qualitative and quantitative methods and original data on the voting behaviour of about 2300 households in 38 villages. The results of this dissertation counter the notions that rural Punjabi voters are dependent and that national elections can be won on the basis of extended kinship networks. My data reveals that the dependence of rural voters that so captivates popular discourse about Pakistani politics describes only about 7 percent of voters, and that kinship networks function more as forums for local collective action than as extended political organisations. I found that a vast majority of rural Punjabi citizens vote as members of village-level vote blocs that are organised by the landed village elite. Nevertheless, most rural Punjabi voters do not participate in vote blocs because of socio-economic dependence. Instead, I found that they are benefitseeking political actors who organise within their kinship networks to strengthen their bargaining position and then give their collective votes to vote bloc leaders who act as broker-patrons and provide access to state officials and services. I also found that voting behaviour varies significantly across villages and across households within the same village. Most of the variation between villages is explained by differences in social structure and varying levels of historical and current land inequality, while the fact that households that lie within the same village behave differently from one another is explained mainly by their wealth and caste status.