Voting for Violence: The Election of Law and Order Politicians in Brazil

• In Event: Justice and Injustice: Political Science Perspectives on Crime and Punishment Mini-Conference Lunchtime Poster Session: New Questions in the Politics of Punishment

Abstract

When do politicians campaigning on a law and order agenda benefit at the polls? Informed by
the classic literature on economic models for welfare preferences, I consider a model that
predicts that voters have redistributive and insurance incentives to vote for law and order
politicians. Redistribution plays a role because the costs from a rise in punitiveness fall mostly
on lowincome
groups; therefore, the affluent have higher benefits and lower costs of voting for
law and order. On the insurance side, I argue that the chance of being a crime victim shapes the
amount of public insurance the voter is willing to buy in the electoral market. In contrast to most
of the economic models of welfare, however, this model predicts that the redistributive and
insurance effects are correlated. In societies where crime has spiked and risk of exposure is high,
my model predicts that the affluent exhibit higher income effects on the support for punitiveness.
The model is empirically tested in the Brazilian case where violence has become a central
concern and politicians from law enforcement agencies have appeared quite frequently on the
ballot. I use finegranted
electoral data aggregated by each voting station for legislative races in
Brazil from 20022018,
and geolocated official data on crime in two major cities in Brazil to test
the model. The results provide a general theory to understand preferences for punitiveness when
voting in the midst of violence.

Author

• Tiago Augusto da Silva Ventura, University of Maryland