Decentralized despotims Subnational incumbents and electoral violence in Nigeria

What explains subnational variation in lethality and electoral violence perpetrators? While elections are regularly held across the Global South, many are marred by violence. Existing research identifies incumbents as the main sponsors of electoral violence, citing their incentive to retain power and their control over state resources. Moreover, studies suggest that incumbents often outsource violence to non-state violence perpetrators to maintain plausible deniability. However, prior work does not sufficiently account for how incumbents organize lethal violence or choose among distinct violence perpetrators. This study addresses that gap by developing and testing a political economy theory that links access to rents, defined as unearned income, to an incumbent’s capacity to sponsor lethal electoral violence. Pitched at the subnational level, the theory argues that rents conditions both the lethality of violence and incumbents’ choice of violence perpetrators. High-rents incumbents, with stronger financial capacity, are more likely to sponsor lethal violence by hiring violence specialists. In contrast, low-rents incumbents, constrained by limited resources, rely on ordinary citizens. The study analyzes the theory through a comparative case study of four Nigerian states, Rivers, Lagos, Plateau, and Nasarawa, during elections held between 2007 and 2019. While all four states operate under the same federal system, findings show that in high-rents Rivers and Lagos, incumbents organize more lethal electoral violence and hire violence specialists such as cult groups and transport unions, respectively, whereas in low-rents Plateau and Nasarawa, incumbents outsource violence to ad hoc ordinary citizen groups. The analysis is based on 126 original in-depth interviews with local politicians, Civil Society organization (CSOs) members, local journalists, and voters, as well as a self-compiled subnational dataset of 1,144 electoral violence events spanning the four electoral cycles. Beyond explaining lethal violence, the dissertation extends the theoretical framework to account for a broader portfolio of campaign strategies. It argues that rents provide politicians’ with the material capacity to fund both violent and non-violent strategies, while broker networks supply the organizational capacity to implement them. Using Afrobarometer survey data (Rounds 5 and 8) from Lagos State, combined with original interviews with party and non-party brokers, including market leaders, transport unions, and ethnic associations, the study finds that broker networks is associated with vote buying and canvassing, while rents predicts coercion, particularly among incumbents. These findings contribute to the literature on electoral violence, campaign strategies, and political organization in weakly institutionalized democracies. More broadly, the dissertation offers a generalizable theory for understanding subnational variation in violent and non-violent electoral campaign strategies in the Global South.