How does the emotional intensity and dispersion within media controversies surrounding a company influence short-term stock market reactions ?

Details

Supervisors

Vande Kerckhove, Corentin

Faculty

Louvain School of Management

Degree label

Master [120] : ingénieur de gestion, à finalité spécialisée

Abstract

enThis thesis investigates how the emotional intensity and dispersion within media controversies surrounding a company relate to short-term stock market reactions, and whether social media adds information beyond traditional news. Taking the firm-day as the unit of analysis, it addresses three gaps in the literature: the near-exclusive reliance of financial studies on aggregated sentiment scores that collapse controversies into a single positive-or-negative value; the resulting neglect of emotional dispersion, the coexistence of divergent reactions, as a potential signal of investor uncertainty in its own right; and the limited use of Large Language Models, rather than lexicon-based tools, to capture the contextual and polarized nature of controversy discourse. Empirically, the study builds a corpus of 3,364 firm-days across 97 S&P 500 companies selected for their exposure to abnormal public attention, identified through Google Trends spikes over 2025. News articles, and, on a pilot scale, posts from X, are scored for sentiment and eight discrete emotions using an LLM, validated against manual annotation, and complemented by lexicon-based measures (VADER, NRC) as a robustness benchmark. Using an event-study design paired with firm-fixed-effects panel regressions, three hypotheses are tested: that emotional intensity drives abnormal returns and volatility, that emotional dispersion drives volatility specifically, and that social media carries incremental information relative to news. A battery of robustness checks: lagged regressors, alternative controversy definitions, lexicon-based replications, and within-firm interaction models, is used to separate genuine signal from artefacts of sample construction. The results are narrower than the initial hypotheses anticipated. Average sentiment shows a modest but robust association with returns, visible across ordinary news days rather than within the controversy episodes it was expected to explain most. Emotional dispersion, despite its theoretical appeal as a marker of disagreement and uncertainty, shows no detectable link to volatility under any specification tested. Social media echoes news sentiment more than it contradicts it, offering a thin but genuine layer of incremental information rather than a distinct emotional regime. Overall, these findings suggest that emotionally structured media coverage relates to financial markets primarily through average tone rather than dispersion, and through returns rather than volatility, a narrower channel than the polarization-and-uncertainty framing common in the controversy literature would predict.