Markets, herding and response to external information

TL;DR

This paper introduces an agent-based market model incorporating herding and external information signals, analyzing how markets respond to external shocks and identifying conditions for optimal information assimilation.

Contribution

It develops a stochastic agent-based model with herding behavior influenced by dynamic external signals, calibrated with real economic sentiment data.

Findings

Market response is maximized at intermediate parameter values.

Three market regimes identified: amplification, accurate assimilation, undervaluation.

Model aligns with real stock market data for calibration.

Abstract

We focus on the influence of external sources of information upon financial markets. In particular, we develop a stochastic agent-based market model characterized by a certain herding behavior as well as allowing traders to be influenced by an external dynamic signal of information. This signal can be interpreted as a time-varying advertising, public perception or rumor, in favor or against one of two possible trading behaviors, thus breaking the symmetry of the system and acting as a continuously varying exogenous shock. As an illustration, we use a well-known German Indicator of Economic Sentiment as information input and compare our results with Germany's leading stock market index, the DAX, in order to calibrate some of the model parameters. We study the conditions for the ensemble of agents to more accurately follow the information input signal. The response of the system to the external information is maximal for an intermediate range of values of a market parameter, suggesting the existence of three different market regimes: amplification, precise assimilation and undervaluation of incoming information.