In emerging democracies, elections are encouraged as a route to democratization. But often, violence and corruption threaten the quality of procedural and substantive democracy delivered through elections. 

What shapes insurgents’ use of electoral violence, and how does violence affect turnout?

Motivated by research that demonstrates the effects of civilian casualties on the ability of insurgents to produce violence in both Iraq (Condra & Shapiro, 2012 (AJPS)) and Afghanistan (Wright et al., 2022), we develop a theory of the strategic considerations that shape insurgents’ use of violence around elections in conflict-ridden democracies. While insurgents want to maximize disruption to elections, they also recognize the need to minimize harm to civilians because such collateral damage increases civilians’ willingness to share information with counterinsurgents. Using geo-referenced data on turnout across polling stations, the ethnic composition and locations of villages, and newly declassified micro-data on insurgent violence, we examine the timing and spatial allocation of insurgent violence and show that patterns of violence across Afghanistan’s elections are consistent with observable implications of our theory. At the micro-level, insurgents’ tactics are remarkably attuned to this strategic constraint on their use of violence. We also estimate the impact of the timing and spatial distribution of insurgent violence on turnout in the 2014 election. We instrument for direct fire attacks and IED placement with surface wind speeds and nighttime cloud cover, respectively, as two environmental constraints on counterinsurgent operations. This strategy shows that violence has a substantial negative effect on turnout, while still minimizing actual harm to civilians. These results are particularly informative for theory because they are based on such a long span of data on insurgent violence around multiple elections within one case, so that we are able to detect and explain patterns over the course of an insurgency, rather than from a snapshot in time. (Condra et al., 2018 (AER))

What are the consequences of safeguarding elections with corrupt police?

Given this threat to elections that comes from violence, governments’ emphasis on security provision around elections to make citizens safer from violence seems prudent, at least in theory. But noting that emerging democracies like Afghanistan are often characterized by security forces that citizens view as extremely corrupt, we took advantage of a unique opportunity to explore how the deployment of corrupt police to protect voters would affect citizen participation in the 2010 Afghan national election. Leveraging our personal experience in the deployment planning process, as well as interviews with officials who made decisions, we estimate police deployments’ impact on turnout using data from the universe of polling sites and various household surveys. Polling centers with similar histories of pre-election violence inadvertently received different deployments of the Afghan National Police, enabling identification of police’s effects on turnout. Increases in police presence decreased voter turnout by an average of 30%. Our results highlight how corrupt security forces can negatively affect government efforts to build legitimacy through elections in weakly institutionalized and conflict-affected states. (Condra et al., 2019 (E&P))