TwinMarket: A Scalable Behavioral and Social Simulation for Financial Markets
Yuzhe Yang, Yifei Zhang, Minghao Wu, Kaidi Zhang, Yunmiao Zhang, Honghai Yu, Yan Hu, Benyou Wang
2025-02-06
Summary
This paper talks about TwinMarket, a new system that uses AI to simulate financial markets by modeling the behavior of individual investors. It shows how personal decisions can lead to larger patterns, like financial bubbles or recessions, helping us better understand how markets work.
What's the problem?
Traditional methods for studying financial markets struggle to capture the complexity of human behavior, especially irrational actions like emotional decisions or cognitive biases. This makes it hard to accurately simulate how individual actions affect the overall market.
What's the solution?
The researchers developed TwinMarket, which uses large language models (LLMs) as virtual agents to simulate investor behavior. These agents mimic real-world decision-making, including emotional and social influences, within a stock market environment. By analyzing how these agents interact, the system reveals how small individual actions can lead to big market changes.
Why it matters?
This research is important because it provides a more realistic way to study financial markets and predict their behavior. It can help policymakers and investors understand how decisions at the individual level impact the economy as a whole, leading to better strategies for managing markets and preventing crises.
Abstract
The study of social emergence has long been a central focus in social science. Traditional modeling approaches, such as rule-based Agent-Based Models (ABMs), struggle to capture the diversity and complexity of human behavior, particularly the irrational factors emphasized in behavioral economics. Recently, large language model (LLM) agents have gained traction as simulation tools for modeling human behavior in social science and role-playing applications. Studies suggest that LLMs can account for cognitive biases, emotional fluctuations, and other non-rational influences, enabling more realistic simulations of socio-economic dynamics. In this work, we introduce TwinMarket, a novel multi-agent framework that leverages LLMs to simulate socio-economic systems. Specifically, we examine how individual behaviors, through interactions and feedback mechanisms, give rise to collective dynamics and emergent phenomena. Through experiments in a simulated stock market environment, we demonstrate how individual actions can trigger group behaviors, leading to emergent outcomes such as financial bubbles and recessions. Our approach provides valuable insights into the complex interplay between individual decision-making and collective socio-economic patterns.