Are Stock Markets Driven More by Sentiments than Efficiency?

Abstract

In this paper, we investigate whether the sentiment contained in text of news items influences a company's stock price. With the means of sentiment analysis, a quantitative approach for measuring sentiment in text, we derive scores for positive and negative sentiment contained in news. Next, we relate these quantities to stock market prices. By regression analysis we show that the sentiment contained in text of news items affects a company's stock prices. This reveals that human sentiments influence capital markets and shows the potential of sentiment analysis as a quantitative approach to measure opinion with possible applications in e.g. public opinion tracking or brand and product management.