War can have long-lasting effects on individual mental health through war trauma. In this paper, we explore the impact of constantly recalling painful episodes related to the 1992-1995 Bosnia and Herzegovina conflict on individual mental health in 2001 using the Center for Epidemiologic Studies Depression (CES-D) scale. Potential endogeneity and reverse causality issues are addressed using objective measures of war intensity recorded at the municipality level. We find that individuals experiencing war trauma have worse mental health six years after the end of the conflict. In particular, instrumental-variable estimates show that they score 16 points (more than 1.5 standard deviations) higher on the CES-D scale (with higher scores meaning more depression symptoms) and have a 60 percentage points higher probability of being at risk of depression. Our results are robust to a number of sensitivity checks accounting for individual geographical mobility and different treatment intensities, and suggest that the negative effects on mental health are not mainly mediated by physical health problems. Back-of-the envelope calculations show large economic costs of war-trauma.