Abstract
The present study intends to assess the impact of three different events that made negative partisanship significant to understand the rise of far-right populist, Jair Bolsonaro: the religious demographic transition driven by Pentecostal evangelicals; the political activism of the Brazilian justice system and its mutualistic relationship with legacy media outlets; and the rise of social media platforms as a performatic space for political communication, exacerbated by automated disinformation devices. In order to estimate the effects of those transformations on voting, two econometric strategies were deployed: beta regression and quantile regression based on aggregated data by municipalities. Additionally, a set of variables based on CETIC microdata were mobilized to correlate the probability of sharing messages through social media and the stratified impacts of disinformation campaigns. The results, obtained through a logit model, compose a parallel analysis to reinforce the results obtained by the estimates of the models for vote proportion.