The influence of social network structure on stock price disclosure

Abstract

We propose a new heterogeneous agent model of a dynamic continuous double auction stock market embedding multiple social network to investigate the influence of different network structures on stock price disclosure. According to the four network formation algorithms (preferential connection, random connection, weighted random connection and random reconnection), we construct scale-free random network, weighted random network and small-world network respectively. We find that the small-world characteristic of social network speed up the diffusion of private information in market, and then improve market efficiency and promote price discovery. Furthermore, to some extent, simulation results indicate that the network structure can explain the occurrence of anomalies like underreaction and overreaction in the financial market.