[Paper Review] Microscopic Models of Financial Markets

Egle Samanidou, Elmar Zschischang+2 more|arXiv (Cornell University)|Oct 17, 2001

Complex Systems and Time Series AnalysisEconomics, Econometrics and Finance153 references22 citations

TL;DR

This paper reviews microscopic models of financial markets developed by economists and physicists, focusing on agent-based models that simulate multi-agent interactions to explain universal statistical features like power-law tails in return distributions and volatility clustering. It examines models such as Lux-Marchesi, Cont-Bouchaud, and Solomon-Levy-Huang, showing how they reproduce stylized facts of financial time series through heterogeneous agent behavior and interaction rules, contributing to a deeper understanding of market dynamics and systemic risk.

ABSTRACT

Submitted to F. Schweitzer (ed.), Microscopic Models for Economic Dynamics, Lecture notes in physics, Springer, Berlin-Heidelberg 2002.kiel.tex

Motivation & Objective

• To examine the development and impact of microscopic models in financial markets from the perspective of both economics and physics.
• To identify and analyze key agent-based models that reproduce universal statistical features of financial time series.
• To bridge the gap between early speculative models of market bubbles and the later recognition of universal scaling laws in financial data.
• To explore how multi-agent systems with heterogeneous agents and interaction rules can replicate empirical stylized facts such as fat-tailed return distributions and volatility clustering.
• To highlight open research questions in modeling financial markets as complex systems of interacting agents.

Proposed method

• Surveying and comparing a range of microscopic models, including Kim-Markowitz, Levy-Levy-Solomon, Cont-Bouchaud, Solomon-Weisbuch, Lux-Marchesi, Donangelo-Sneppen, and Solomon-Levy-Huang.
• Analyzing the Donangelo-Sneppen model of monetary exchange as a foundational model for agent-based wealth dynamics.
• Evaluating models based on their ability to reproduce empirical regularities such as power-law tails in return distributions (with exponent ~3) and temporal scaling of volatility.
• Using simulation-based approaches to study emergent market behavior from local agent interactions, emphasizing agent heterogeneity and herding effects.
• Comparing theoretical predictions of models with empirical financial time series data to validate their explanatory power.
• Framing financial markets as complex systems of interacting agents, drawing analogies to statistical physics and out-of-equilibrium systems.

Experimental results

Research questions

• RQ1 How do microscopic agent-based models reproduce the universal statistical features of financial time series, such as fat-tailed return distributions?
• RQ2 What mechanisms in agent interactions lead to volatility clustering and power-law scaling in financial returns?
• RQ3 How do models like Lux-Marchesi and Cont-Bouchaud explain market crashes and bubbles through agent behavior and herding?
• RQ4 What is the role of agent heterogeneity and interaction rules in generating emergent market dynamics?
• RQ5 What are the open challenges in modeling financial markets as complex systems of interacting agents?

Key findings

• The Donangelo-Sneppen model of monetary exchange provides a foundational framework for understanding wealth distribution and agent interactions in financial systems.
• Models such as Cont-Bouchaud and Lux-Marchesi successfully reproduce power-law tails in return distributions with an exponent close to three, matching empirical observations.
• Volatility clustering and temporal scaling in financial time series are emergent properties of agent-based models with heterogeneous agents and herding behavior.
• The transition from early models focused on market bubbles to later models incorporating universal scaling laws reflects a growing understanding of financial market dynamics as complex systems.
• Agent-based models with simple interaction rules can generate complex market phenomena such as crashes and booms without assuming rational expectations.
• The convergence of insights from physics and economics in modeling financial markets has led to a more robust framework for understanding systemic risk and market instability.