Trump Trading: Quantifying the Impact of Social Media Posts from Public Figures on Intraday Trading

Abstract

This thesis attempts to examine public figures and the impact of their social media activity on intraday stock market returns and volatility. While past literature documents the vast influence of Donald Trump’s posts, it is unknown whether these effects are unique or part of a broader phenomenon. To address this gap, the thesis analyzes posts from Trump, Elon Musk, and Joe Biden, looking into their relationship with short-horizon market movements. Using Twitter post data matched with high-frequency intraday price data from WRDS for SPY and QQQ, I construct intraday event windows and compute returns and volatility. Sentiment is measured using VADER and FinBERT, and engagement metrics and keyword indicators are included. Regression analysis estimates the effects of sentiment, engagement, keywords, and author identity on market behavior. The results show limited evidence of systematic differences across figures at short time horizons, with coefficients mainly small and statistically insignificant. However, engagement appears to amplify the impact of Musk’s posts. For Trump, higher pre-event volatility is associated with more negative returns thereafter, while lower volatility is linked to more positive returns, suggesting a reversal-type effect. Overall, the results indicate that social media activity can have modest effects on market fluctuations, but these effects appear not uniquely attributable to any single individual, reflecting a broader, social media-driven market phenomenon.