Combining Empirical Data with Multi Agent Social Simulation: Investigating Micro-Macro Links in Stock Markets

Abstract

This paper investigates micro-macro links in stock markets. Empirical investigations on individual investors' decision-making and interactions are used to formalize agent rules for an artificial stock market. Simulation of the social interactions of multiple agents in this market generates macro results, like stock market prices and returns over time. The outcomes of the artificial market are finally compared to empirical facts from real stock markets. Taken together, the paper is the first example of an Axtell & Epstein level 3 model in financial markets.

Keywords: multi-agent simulation, artificial stock markets, stock market characteristics, micro-macro links, behavioral finance

JEL Classification: D40, D79, D83, G10, Z13