Social interaction, volatility clustering, and momentum

This paper incorporates information uncertainty and social interaction among investors into a random utility framework and develops a dynamic equilibrium model of asset pricing and investor choice. We show that strong social interaction can lead to endogenous switching between two persistent regimes for the mean choice fraction of investor population, which can simultaneously generate volatility clustering and time-series momentum in asset returns. By using StockTwits post volume as a proxy for social interaction, we provide empirical evidence for the model predictions for various equity indices.