Social Media as a Bank Run Catalyst
Abstract
After the run on Silicon Valley Bank (SVB) in March 2023, U.S. regional banks entered a period of significant distress. We quantify social media's role in this distress using comprehensive Twitter data. During the SVB run period, banks with high pre-existing exposure to Twitter lost 4.3 percentage points more stock market value. Moreover, Twitter pre-exposure interacts significantly with classical run risks to predict greater run severity and greater deposit outflows during Q1-2023, effects unexplained by other banking or market characteristics. At the hourly frequency during the run, high Twitter attention over the past four hours predicts stock market losses, especially for banks with high run risks. By contrast, we find that negative Twitter sentiment does not amplify bank run risks. Rather, our evidence points to a distinctive role of Twitter attention, particularly when tweets are retweeted broadly.