Prefeitos, Platforms, and the Police: Public Security in Brazilian Elections 2012-2022

How and when do Brazilian politicians prioritize public security in electoral campaigns? I propose that an increasing number of police officers running for office have forced other candidates to discuss crime reduction policies more frequently and creatively in campaign materials. This is particularly true of local politicians, partly explaining trends toward municipalization of public security in Brazil. This project will analyze a corpus of over 60,000 executive-level platforms, systematically identifying when and to what degree politicians discuss locally important policy issues, including public security. Kellogg funds will be used to hire a team of three research assistants to manually code a sample of these platforms at the sentence level. These manual codings will then be used to train \textit{supervised} machine learning models to automatically read and label the remaining platforms. The output of these models will be used in "traditional" statistical analysis to test the implications of my theoretical framework.