Computational Platforms for Agent-Based Financial Models

This research project constructs a broad set of software tools designed to better facilitate the understanding and comparative features of various types of agent-based finical markets.

Agent-based computer simulations are a fundamental tool for understanding financial market dynamics. These stylized computer thought experiments provide a view of markets as systems of interacting, learning, and evolving agents. They have given us a picture of what is happening in markets far from equilibrium and have made progress in explaining empirical puzzles we see in financial time series. They also remind us that the common assumptions used in equilibrium financial market models form a straight jacket that has kept much traditional economic thought locked in a limited domain. Though computer simulated markets have been with us for over 25 years, common software tools are still not available for researchers interested in getting started in this area. This project builds a set of software tools which can be used by researchers and others interested in exploring the fields for the first time or as part of a curriculum in agent-based economic modeling. The tools also form a comparative theoretical test bed to better understand what the critical theoretical features are and how they impact our broader understanding of actual market dynamics.