Why factions matter : a theory of party dominance at the subnational level

What explains the resilience of formerly nationally dominant parties at the subnational level? This dissertation demonstrates that factionalism is key. When intra-party factions are united, subnational dominant parties retain power even under adverse electoral conditions. By contrast, divisions and conflicts among internal groups lead these parties to lose even in favorable electoral contexts. I test these claims using a variety of quantitative and qualitative evidence from Mexico, focusing on the electoral performance of the Revolutionary Institutional Party (PRI) in contemporary gubernatorial elections. Democratization potentially undermines unity in dominant parties because it provides politicians with viable exit options (i.e., joining the opposition) and because authoritarian central party committees no longer control subnational politics. Yet, I argue that factions can cooperate under democracy when they were more autonomous from the center during the authoritarian period. The negotiation skills acquired in the past help them "get along" in the absence of an external enforcer. By contrast, previously subordinated factions never acquired such skills and quickly became antagonistic to each other under democracy. As I show, collaboration had positive electoral consequences in subnational elections whereas antagonism had pernicious ones.