Competing Narratives in Action : An Empirical Analysis of Model Adoption Dynamics

We use a longitudinal dataset measuring beliefs and behaviors to study the dynamics of model - or narrative - adoption during the Covid-19 pandemic. We show that individuals switch beliefs about the effectiveness of preventive behaviors following changes in perceived risk. The adoption of narratives promoting preventive behaviors is procyclical and narrative switching is influenced by exposure to conflicting information. We explain the data using a heterogeneous-agent model of competing narratives in which agents exhibit motivated beliefs. Adopting misspecified narratives increases infection rates, highlighting the importance of promoting accurate beliefs to guide behavior in the presence of novel risks