Social media, political uncertainty, and stock markets

Research output: Contribution to journalArticle

Authors

External organisations

  • Swansea University
  • University of Reading

Abstract

This study proposes a new measure of firm-level uncertainty exposure around important political events. More specifically, we construct a degree of (dis)agreement among social media users who jointly mention firms and politicians. We study a sample of over 23 million tweets mentioning both a firm from the S&P 500 composite and ‘Trump’ from October 2016 to May 2017. We then analyze the relationship between the (dis)agreement measure and individual stock features. The results suggest that increased disagreement among such tweets is associated with heightened stock price volatility and trading volume. This link is observed before the US Presidential Inauguration in January 2017 but not afterwards. The finding is confirmed by further robustness checks based on filtered tweets with policy keywords and policy-sensitive industries.

Details

Original languageEnglish
JournalReview of Quantitative Finance and Accounting
Early online date21 Jan 2020
Publication statusE-pub ahead of print - 21 Jan 2020

Keywords

  • Twitter, investor sentiment, text classification, US election, stock market, computational linguistics