Does Sentiment Analysis of Financial News Predict Stock Price Movements in Emerging Markets

Abstract

The study explores how sentiment analysis of financial news can be used to predict the stock price movement in the emerging markets. The study aims to fill a gap in the literature by considering Brazil (Bovespa), India (Nifty 50), and China (Shanghai Composite) between 2020 and 2025 and using available tools to test the theory on volatile and retail-based markets where behavioural bias is high. A quantitative method was employed to use the VADER sentiment algorithm on 512 English language news items in Reuters and Bloomberg to obtain daily polarity scores (-1 to +1). These scores were correlated with next day stock returns and were analysed using Pearson correlation and Ordinary Least Squares (OLS) regression. The processing of primary data was in the form of structured NLP pipeline, which was transparent and reproducible. Findings show that news sentiment and returns have positive statistically significant relationship in all markets. The variance in the returns was attributed to sentiment to the tune of 14.4 percent in India, 11.6 percent in China and 8.4 percent in Brazil (pooled R 2 = 0.109). India was found to be the most predictive and Brazil was found more sensitive to external shocks. In big crises, anomalies were observed, which shows the weakness of sentiment. The research finds that sentiment analysis provides modest yet useful value of prediction in emerging markets. Results confirm the theory of behavioural finance and give the investors a convenient additional decision-making tool in the short term.