Armed Political Parties and Their Violence: Evidence from West Bengal, India

• In Event: Learning about Political Violence through Observational Data

Abstract

In myriad democracies, armed political parties directly perpetrate violence against political opponents. However, existing theory is ambiguous about why they perpetrate pre- and post-election attacks against opposition parties in some localities, but not others, within the same electoral context. In this study, I contend that a locally-elected representative's partisan alignment with the government's ruling party increases its incentives (i.e., potential payoff) and ability (e.g., access to material resources) to perpetrate pre- and post-election attacks against partisan opponents. Violence is intended to consolidate territorial control, which minimizes the electoral threat posed by other parties, and ensures the incumbent's ability to implement policy without encountering local resistance. The empirical test of this theory is twofold. First, I implement a regression discontinuity design using a novel dataset of political party violence around state assembly elections (1982—2016) in the Indian state of West Bengal. Additionally, I evaluate mechanisms by drawing upon historical sources and extensive fieldwork which involved over fifty interviews with local political party workers and leaders in West Bengal. I discuss the implications of this study for theories of democracy and violence as well as the external validity of these findings.

Author

• Drew Stommes, Yale University