Taxing Dissent: The Impact of a Social Media Tax in Uganda

Abstract

We examine the impact of a new tool for digital repression—a daily tax on social media use in Uganda. Using a synthetic control framework and exploiting the exogenous timing of the tax induced by the legislative calendar, we estimate that the tax reduced the number of georeferenced Twitter users by 14 percent. The effects are larger for poorer and less frequent users. Despite the overall decline in Twitter use, tweets referencing collective action and observed protests both increased around the onset of the tax relative to the synthetic control. The high salience of the tax as digital repression and its impact on the composition of users are two potential mechanisms for this backlash effect.

Keywords: social media, censorship, political dissent, protest, riot, tax

JEL Classification: D72, P52