ESG News Sentiment and Stock Price Reactions: A Comprehensive Investigation via BERT

How much do equity prices actually move when a sustainability headline breaks out?

Gregor Dorfleitner and Rongxin Zhang study short-horizon market responses to sustainability news sentiment in their paper « ESG News Sentiment and Stock Price Reactions: A Comprehensive Investigation via BERT ».

They explore whether investors react asymmetrically to positive versus negative sustainability news, and whether a firm’s historical ESG score moderates these reactions based on Thomson Reuters’ Eikon news from May 2019 to March 2021.

They show that:

• They can rely on 84,835 news items across 13,327 stocks: the final sentiment mix is 51% neutral, 44% positive, and 5% negative.
• On the event day, average abnormal returns are +0.31% for positive sustainability news and -0.75% for negative ESG ones, both statistically significant.
• Over the 1-day window around publication, positive news earns +1.17% while negative news loses -1.28%. Positive gains persist at longer horizons (+1.24% after 10 days).
• In America, event-day abnormal returns are +0.37% for positive news and -1.01% for negative news, with +1.38% versus -2.10% after 1 day, while Europe shows +0.34% versus -0.78% on the day and +1.16% versus -1.82% after 1 day.
• Positive ESG news is followed by price drift consistent with underreaction, while negative ESG news is followed by reversal consistent with overreaction.
• Higher ESG scores soften the price hit from negative sustainability news, while positive sustainability news triggers larger gains for firms with worse ESG records.

This research suggests investors should treat ESG headlines as measurable short-term risk shocks, not just narrative noise, and explicitly condition those shocks on a firm’s sustainability track record.

Firms with weaker records can benefit more from positive sustainability news, with effects spanning beyond the announcement day as the market integrates the information.

As a limitation, the authors note provider opacity can make results provider-dependent, and the large share of positive ESG news may include many items that do not trigger strong market reactions and potentially mask « high-signal » positive events.

Similarly, the analysis does not differentiate sustainability news types by financial materiality or verify authenticity, even though the authors discuss exaggeration as a possible interpretation.