The impact of political assassinations on turnout: Evidence from Colombia

Abstract

Although a growing literature has investigated the effects of various types of civil war violence on political behavior, no study has examined the impact of assassinations targeting politicians. This is a critical omission, as violence against local politicians is prevalent across civil war contexts and may be the most consequential form of violence for political participation by affecting both candidate supply and voter demand. Using an original dataset of nearly 2,000 killings of Colombian local politicians between 1980 and 2023, we estimate the impact of this violence on voter turnout. Taking municipalities where assassination attempts failed as a comparison group, we find that political assassinations significantly decrease voter turnout in both the short and medium terms, with effects persisting in various elections even after the signing of a peace agreement. These findings contrast with many studies suggesting that other forms of civil war violence enhance political participation during the postconflict period or after a truce or peace agreement. Our results suggest that different forms of violence can have distinct effects on political behavior, underscoring the need to theorize how the targeting, nature, and context of violence condition its effects. This echoes calls for more nuanced studies on the behavioral impacts of violence. Our findings also have implications for understanding democracy amid rising violence against political leaders in countries affected by organized crime, such as Mexico and Brazil; polarized contexts, such as the United States; and weakly institutionalized democracies, such as South Africa, Indonesia, and the Philippines.