Assessing the impact of news sentiment on stock returns in the Russian market

Abstract

Subject. The relationship between news sentiment and securities returns. Objectives. To quantitatively assess the impact of news sentiment on the returns of stocks in the Russian stock market and to identify the asymmetry in price reactions to positive and negative information signals, taking into account time horizons, issuer market capitalization, and industry affiliation. Methods. The study employed sentiment analysis methods for text data based on transformer deep learning models, as well as the event‑study methodology. The empirical base includes minute‑level data for the stocks of the 50 largest issuers on the Moscow Exchange and a corpus of messages from specialized Telegram channels for the period 2020–2025. To test the hypotheses, Welch’s t‑test and regression models with robust standard errors were applied. Results. A stable asymmetry in market reaction was identified: negative news causes more pronounced price changes compared to positive news across all analyzed time horizons, with the largest effect observed in the short‑term interval. Overall, no dependence of return sensitivity to sentiment on issuer market capitalization was found. A cross‑sectoral effect of news from large companies on the returns of smaller issuers was detected over medium‑term and daily horizons. Conclusions. News sentiment is a significant factor in the short‑term dynamics of stocks in the Russian stock market. The obtained results enhance the understanding of pricing mechanisms in emerging markets and can be used to develop quantitative investment strategies that account for reactions to information signals.