How does political violence affect voting behavior? In order to answer this question, I examine the impact of fatalities in Turkey's PKK conflict between 2015 and 2018 on voting behavior in Turkey in both parliamentary and local elections. Using an original, district-level dataset on the location of state security forces, civilian and insurgent fatalities, I study how both the presence and the number of fatalities affect support for political parties: I focus on the incumbent Justice and Development Party (AKP), the Turkish nationalist action party (MHP), and the Kurdish party (HDP). I find that fatalities increase the vote share of the Turkish nationalist party and lowers the support for the opposition Kurdish party. I also find mild increase in support for the incumbent JDP in conflict-affected districts. The robustness checks show that these findings can't be explained only by potential irregularities in the counting of votes, the deployment of nationalist security forces and their families to the conflict-affected areas, and out-migration of particular groups from conflict-affected areas. Overall, these findings contribute to a growing body of research that show that conflict can increase the vote share of right-wing, nationalist parties while it lowers public support for parties that are aligned with the political goals of insurgents.