[2Yin-B] Poster 4

[2Yin-B-36] Corporate Valuation Using an Artificial Market Composed of Multiple Types of Investor Agents

〇Yuki Tsuboi¹, Kunihiko Harada¹, Hiroaki Ozaki¹ (1. Hitachi, Ltd.)

Keywords:

Artificial Market,Multi-Agent Simulation,Corporate Valuation,Agent-based model

This paper reports on corporate valuation using an artificial market composed of multiple types of investor agents to identify corporate value determinants. Recently, non-financial indicators have gained importance alongside financial ones. However, integrating these diverse indicators into a unified framework is challenging due to differences in scale and dynamics. Moreover, price formation mechanisms in markets with diverse participants are complex, and existing methods lack the explanatory power to establish high-resolution evaluation metrics. To address this, we construct a market of multiple types of investor agents trading based on various corporate indicators and propose a method that estimates investor composition so that stock price trends are calibrated to real data. While primarily aimed at explaining market phenomena, validation using historical data confirmed that this method outperforms conventional statistical methods in reproducing stock price fluctuations. This approach enables a quantitative explanation of value formation by decomposing factors according to investor tendencies and corporate indicators.