Topology and Behaviour of Agents: Capital Markets

TL;DR

This paper models agent interactions in capital markets, focusing on information exchange, risk aversion, and market stability, revealing how social behaviors influence investment strategies and market states.

Contribution

It introduces a novel model linking social interaction, information verification, and risk aversion to market stability and agent behavior in capital markets.

Findings

Optimal information exchange balances expected returns and risk.

Market stability depends on parameters like expected return and risk.

Variance of subgroup distributions relates to market dynamics.

Abstract

On a capital market the social group is formed from traders. Individual behaviour of agents is influenced by the need to associate with other agents and to obtain the approval of other agents in the group. Making decisions an individual equates own needs with those of the other agents. Any two agents from the group may interact. The interaction consists of the exchange of information and it costs some money. We assume that agents give reference to the origin of the information if asked by other agents. Thus the agent may verify obtained private information. Hudak recently used methods described by Rivier to study social behaviour of such agents. He characterized the quantity which corresponds to verification of information. Quantity which characterizes verification of information contributes to an aversion of an agent with respect to a risk. The mix of investments of an agent in a given cell with an average measure A of risk aversion in the cell is found from minimum of the average per cell aim function $<FM>$. Absolute minimum corresponds to such a state in which there is an optimal mix of the exchange of information for a given expectations about the capital market. The crowd and personal /$\approx <f>$/ contributions to the risk aversion of an agent are present in the aversion constant A. We have discussed a stable and metastable states of the market for different values of E, an expected return for a given investment period, of EV, an expected risk for a given investment period, and of b, a constant which characterizes contribution of the quantity $<f>$ to the risk aversion. Variance of n for the distribution of nonreducibile subgroups is found. Our model describes intermediary process effects.