Terrorist Violence and Incumbent Support: Case of Turkey

• In Event: Innovations in the Empirical Study of Terrorism

Abstract

Terrorism often seeks to impact democratic politics, and terrorist violence does impinge on electoral landscape but it is not clear whether it benefits or harms the incumbent party. Identifying the effect of violence is difficult because researchers cannot manipulate timing of the attacks and elections are only held once in every few years. Incumbency and party composition of the government are hard to disentangle in some contexts in the presence of coalition governments, thus findings from single case studies may be capturing changes in right-wing support rather than incumbency. Furthermore, average nationwide effects are likely to mask subnational variations in effects of violence, which may not be compatible with the national effects. There is indeed a dearth of studies that examine how geographic variation in manifestation of violence affect party vote shares in an election. To address all these issues, this project examines the case of Turkey— a conflict-ridden context beset by both domestic and international terrorism, and that has had a single-party government between 2002-2018. I use three distinct datasets to explore the effect of terrorist events on electoral landscape. First dataset is pooled time-series crosssectional survey data gathered monthly from a representative sample between 2010 and 2017. The second dataset is an original dataset that has district-level terrorist events and casualties between 2000 and 2018, coded both by location of death and burial. The third dataset is the official district-level election results from the five elections that occurred between 2002 and 2018.

In this paper, I focus on experiences across different geographies and argue that the local manifestations of violence play an important role in how violence, terrorist or rebel attacks, which are at times crosscutting, reflects on party vote shares. In particular, I discuss direct and indirect experiences with violence. While direct refers to active clashes and violent conflict in a locality that poses a direct security threat, indirect violence refers to signal of violence such as soldier funerals without the direct threat. Drawing on this distinction and utilizing the new datasets, this paper will offer a thorough account of how terrorism affects vote shares of the incumbent party with the most detailed dataset available to date from Turkey.

Author

• Sule Yaylaci, Georgetown University