Listening to the Buzz: Social Media Sentiment and Retail Depositors' Trust

Abstract

We investigate the relationship between the rumours on Twitter regarding banks and deposits growth. The sentiment expressed in tweets is analysed and employed for the nowcasting of retail deposits. We show that a Twitter-based indicator of sentiment improves the predictions of a standard benchmark model of depositor discipline based on financial data. We further improve the power of the model introducing a Twitter-based indicator of perceived interconnection, that takes into account spillover effects across banks.