A social-media-based approach to predicting stock comovement

Abstract

Stock return comovement analysis is important to financial analysts, decision makers, and academic researchers and has many financial implications, such as portfolio management, style investing, and market risk detecting. This paper proposes a novel model to both identify homogeneous stock groups and predict stock comovement with respect to firm-specific social media metrics. One of the innovations of the social media platform is that it breaks traditional media intermediation. A firm with an official Twitter account can publish information and interact with its users directly. Such direct information is largely reflected on firm-specific metrics, e.g., the firm’s number of followers and number of tweets sent. To the best of our knowledge, this paper is the first to reveal the impact of social media metrics on stock return comovement studies. By analyzing samples from the NYSE and NASDAQ stock exchanges, we find that firms with official Twitter accounts have a much higher comovement than those without such accounts. Furthermore, we classify the former set of firms into homogeneous groups by their specific microblogging metrics. The results demonstrate that these metrics cannot only predict the comovement of stocks but also notably increase the accuracy of comovement predicting, compared with industry categories.