In this paper, we empirically examine the social consequences of electoral violence with a focus on the first mayoral elections in post-democratization Indonesia, which started in 2005. During these elections, voters’ dissatisfaction with the poor quality of electoral administration induced violent incidents in many municipalities. Using the nationwide household survey administered in 2006, we estimate the impacts of holding an election, which was typically accompanied by violence, on voters’ social attitudes. The identification strategy in making causal inference is to leverage a variation in the election timing that is as-if random for historical reasons. Using 40 pre-determined covariates measured at the level of municipalities, we show that our assumption of as-if randomness is strongly supported by data. We also conduct the exact matching to secure the balance of a range of respondent-level pre-determined variables. The results of causal analysis show that respondents who experienced the first mayoral election before the survey was administered tend to distrust neighbors and to participate less in community programs. We conclude that democratization and positive social changes are not necessarily the two sides of the same coin. Democratization without electoral integrity is likely to produce unintended negative social outcomes.