Do Donald Trump's tweets reduce the market relevance of official U.S. economic announcements?
Files
Duthoit_60572100_2026.pdf
Open access - Adobe PDF
- 1.24 MB
Details
- Supervisors
- Faculty
- Degree label
- Abstract
- This thesis examines whether and how Donald Trump's social media activity affects U.S. financial markets. We use a dataset covering the full 2016 to 2025 period across two presidential terms. Three hypotheses were tested on daily S&P 500 returns and changes in two-year and ten-year Treasury yields. Firstly, H1 asks whether tweet volume increases conditional market volatility. Secondly, H2 describes whether pre-announcement tweet activity reduces the market-moving capacity of scheduled macroeconomic releases. And thirdly, H3 asks whether negatively framed tweets generate larger volatility responses than positive ones. H1 and H3 are estimated using GJR-GARCH(1,1) models on announcement-free trading days. H2 is tested through an event study framework covering 252 FOMC, NFP, and CPI announcements, estimated by OLS with HC3 robust standard errors. Five sets of robustness checks are reported. The results are mixed but coherent. H1 receives marginal support on the S&P 500 under QML inference only. The signal strengthens to 1% significance when the analysis is restricted to financially relevant tweets. It is consistent with attenuation bias in the baseline. H2 is the most robust finding, confirmed at 5% and stable across all window specifications. H3 is not confirmed under any specification. Overall, the evidence suggests that Trump’s social media communication affects financial markets primarily through volume and attention saturation, not through the emotional content of individual posts. How much Trump tweets matters more than what he says.