Spatial Dynamics of Election Violence: How Repression Spreads Dissent around Elections

How does political violence spread around election times within countries? Although election times are periods possibly most vulnerable to the contagion of violence, we know very little about how election violence spreads spatially. We argue that government-sponsored election violence in one area will increase the levels of election violence inflicted by antisystemic actors in another area with similar political or socioeconomic characteristics. Government violence in one area increases the expectation that repression would soon start in another area that looks similar to the targeted area, prompting opposition groups in these areas to take action preemptively. We test our arguments using subnational data on India's election violence from 1991 to 2009, finding that government election violence in one state increases opposition-led election violence in another state with similar political and socioeconomic characteristics in India. Our results are robust when using a stationary causal directed acyclic graph approach and recover the bias-corrected spatial effects.