The Structure of Twitter Cashtag Networks and Market Reaction to Earnings News

The aim of the Essex Accounting Centre (EAC) research seminar series is to support our world-class research activities in five key areas: accounting and global development; capital Markets, audit, regulation & reporting; publicness and resilience, precarity, exclusion & social justice; and environment, climate change & vulnerability. The seminar series is also expected to promote interdisciplinary research that links the work of members of the centre with others both within the university and with external institutions.

Networks pervade our social and professional lives, ranging from golfing partnerships and lunch ties to inter-locking board-of-director relations, Facebook friendships, and LinkedIn connections.

A core premise of the field of social network analysis is that information flows are heavily influenced by network structure, including the extent to which the network is dominated by a few key actors, how inter-connected network members are, and the prevalence of sub-groups and cliques. We apply this logic to the stock market, analyzing the ego network structure of Twitter cashtag discussion networks for stocks in the S&P 1,500 index.

We posit that the structure of these cashtag networks – specifically, their centralization, density, clustering, and prevalence of isolates – affects market reactions to quarterly earnings announcements. We find that abnormal market returns are negatively associated with the degree of centralization and density and positively associated with the extent of clustered sub-groups and the prevalence of unconnected isolates. These results highlight the importance of understanding network structures in predicting stock market movements, offering novel insights into the behavioral aspects of financial markets.