The aim of my thesis is to explain the simultaneous consolidation of clientelism and democracy in developing countries. Democracy has created new spaces for representation and political accountability, but it has also created incentives for maintaining and nurturing clientelistic bonds. Why some political parties choose clientelistic inducements to mobilize voters, and why some clientelistic parties succeed in mobilizing low-income voters while others fail, are the questions I seek to answer in my dissertation. Whereas these questions are motivated largely by puzzles in the consolidation of democracy in Latin America, these are questions of importance outside the region and the study of machine politics.

My dissertation argues that the combination of broker access and agency in distributing clientelistic inducements, their capacity to build networks to trade goods and favors for votes, and their strategic choices in deciding how to mobilize voters explains variation in the dependent variable. By combining field and archival research, participant observation, and in-depth interviews with party bosses, brokers, activists, community organizers, and voters across provinces, municipalities, and neighborhoods in Argentina, my thesis shows how informal incentives explain the persistence and consolidation of clientelistic strategies in political districts inhabited by low-income voters. I find that when political parties choose to reward and punish candidates or brokers, terms that I use interchangeably, only based on how many voters they can mobilize, brokers competing for the support of low-income voters are encouraged to distribute clientelistic inducements and monitor voters. I also find that party brokers who employ coercive strategies, such as threatening voters to withdraw benefits if voters fail to turn out, are more effective than those that choose to employ persuasive strategies.

By examining and comparing political party organization across and within subunits, I find that when faced with the same incentives, candidate strategic choices are not affected by their partisanship. In addition, through "thick" description, I show how brokers systematically rotate state aid among voters and promise future benefits in exchange for actual support. As a result of these tactics, clientelism influences the behavior of a broader set of voters than those who receive immediate benefits before elections.