Ernesto Calvo, University of Maryland

"A model of voting with motivated reasoners: with applications to affect polarization"


What is the optimal policy response to an angry voter? It has been extensively documented that voters’ positive and negative feelings about parties and candidates shape how they place them on the ideological space. This is the result of motivated reasoning, where voters seek out information that aligns with their “likes” and “dislikes” and use this information to infer the position of parties. In this article, we leverage these well documented biases to describe the effect of non-policy shocks, such as an increase in affect polarization or a COVID-19 diagnostic, on the optimal ideological placement of parties. More generally, we derive an equilibrium model of voting under motivated reasoning (i.e. a model for motivated downsian voters). The proposed model is extremely general and accommodates known survey evidence on perceptual bias. We exemplify its usefulness by analyzing 50 years of presidential elections in the US and reanalyzing experimental data on negative social media messages.