The Effects of Online Content Moderation: Evidence from President Trump's Account Deletion

Abstract

We study the effects of online content moderation on user behavior in the context of a prominent case study: the deletion of President Donald Trump's Twitter account on January 8th, 2021. We provide four key findings. First, the toxicity of tweets sent by Trump followers relative to a representative sample of US Twitter users dropped by around 25% after the account deletion. Second, this effect is larger for Republicans and non-political users. Third, Trump's suspension reduced the total number of tweets, suggesting a drop in engagement. Fourth, we find effects on individuals who did not follow Trump directly but followed somebody that did, suggesting network spillovers. This evidence shows that removing a prominent, polarizing individual can reduce the toxicity of online discourse.

Keywords: social media, content moderation, hate speech