Title: Forecasting Electoral Violence
Authors: David Randahl, Maxine Leis, Tim Gåsste, Hanne Fjelde, Håvard Hegre, Staffan I. Lindberg, and Steven Wilson
Issue: V-Dem Working Paper, Series 2024:150
Publisher: The Varieties of Democracy Institute, University of Gothenburg
Abstract: Electoral violence remains a significant challenge worldwide. It not only threatens to undermine the legitimacy and fairness of electoral outcomes, but often has serious repercussions on political stability more broadly. The ability to prevent electoral violence is critical for safeguarding democracy and ensuring peaceful transitions of political power. Predicting which elections are at risk of violence is an important step for effective prevention. In this study, we build and train a set of machine-learning models to forecast the likelihood of electoral violence on a global scale. Using a comprehensive set of data sources, with features including economic indicators, records of historical violence, political instability, and digital vulnerability, we predict the risk of electoral violence on a scale from no violence to severe violence. When combining a subset of these models to produce ensemble predictions of electoral violence for 2024-2025, our results show that our model effectively discriminates between the different levels of risk with a high degree of predictive accuracy. This research contributes to the field of political violence prediction by providing a medium-term data-driven forecasting tool for electoral violence. This knowledge may assist practitioners in the field of violence prevention by pinpointing elections at risk.
