Artificial Markets

Fully electronic market will be ubiquitous in the near future. Because financial markets are the most efficient and best studied of all markets, they can provide unique insights in designing the next generation of electronic markets. In particular, in addition to automated electronic financial markets, there will be similar markets for bandwidth, for telephone time, and for many other commodities. The electronic markets of the future will achieve real-time, efficient and transparent allocation of resources between people and organizations and within electronic networks. In this project, we propose to study computational systems of loosely coupled, asynchronous, adaptable software agents with learning abilities. We will design, implement, and characterize artificial markets in which software agents endowed with different learning modules can interact, evolve, and compete.

Relevant Publications and Preprints:

• Chan, N., LeBaron, B., Lo, A., and T. Poggio, 2001, "Agent-Based Models of Financial Markets: A Comparison with Experimental Markets," in revision for Journal of Financial Markets.
  
• Hutchinson, J., Lo, A., and T. Poggio, 1994, "A Nonparametric Approach to Pricing and Hedging Derivative Securities via Learning Networks'', Journal of Finance 49, 851--889.
  
• Lo, A., "Neural Networks and Other Nonparametric Techniques in Economics and Finance'', in H. Russell Fogler, ed.: Blending Quantitative and Traditional Equity Analysis, 1994. Charlottesville, VA: Association for Investment Management and Research.