Prof. Dr. Thomas Lux

 

Chair

Institut für Volkswirtschaftslehre Universität Kiel
Olshausenstraße 40
D-24118 Kiel

e-mail: lux@bwl.uni-kiel.de

Web-site: http://www.bwl.uni-kiel.de/vwlinstitute/gwrp/team/lux.htm

The current research interests of Prof. Lux and his research group focus on various theoretical and empirical aspects of financial markets and monetary economics. Their methological approach is a very broad one attempting to combine standard tools of economics and econometrics with new approaches for multi-agent modeling adopted from physics and computer science. Thomas Lux published one of the first papers on applications of statistical mechanics in a finance setting in 1995 in the Economic Journal and has elaborated on this subject in his subsequent research. Together with Michele Marchesi, he is co-author of a widely quoted paper on microscopic stock market simulations published in Nature 397, 1999. In this paper, it could be shown that some of the universal properties of financial data (fat tails of returns and clustering of volatility) can be generated as emergent phenomena from the interaction of heterogeneous agents in a prototype artificial market. The Lux-Marchesi model extends earlier behavioral models by treating agents' behavior in a probabilistic manner which allows formal analysis of a large ensemble of autonomous agents via methods borrowed from statistical physics (e.g., mean-field approximations, Master equations).The approach of this and related papers combines facets of the research areas of behavioral finance , agent-based computational economics and ‘econophysics' . In particular, the modeling of interactions of different trader types is inspired by the empirical failure of the representative agent methodology (which appeared unable to explain phenomena like price bubbles and crashes as well as the above empirical regulatories) and the development of behavioral models of trading in financial markets. Further elaborations on this topic have been published in field journals in both economics and physics. This ongoing research also includes modeling of financial markets using artificial intelligence methods (genetic algorithms, genetic programming, classifier systems) as learning algorithms of autonomous agents as well as research on multi-fractal cascades (first introduced in the statistical physics literature on turbulent flows) as models of financial volatility. Thomas Lux has furthermore worked on economic applications of multiplicative stochastic processes with physicist Didier Sornette and he is currently working together with an international group of sociologists and political scientists on a book project on ‘Power Laws in the Social Sciences'.

 

Some Selected Publications:

‘Genetic Algorithms as an Explanation of the Stylized Facts of Foreign Exchange Markets' (with S. Schornstein), Journal of Mathematical Economics (in press)

‘On Rational Speculative Bubbles and Fat Tails' (with D. Sornette), in: Journal of Money, Credit, and Banking 34 (2002), pp. 589 - 610

‘Scaling and Criticality in a Stochastic Multi-Agent Model of a Financial Market' (with M. Marchesi), in: Nature 397 (1999), pp. 498 – 500

‘Herd Behaviour, Bubbles and Crashes', in: Economic Journal 105 (1995), pp. 881 - 896