Research Interests
I have broad interests in complex systems, and during my career I have worked in many areas, including dynamical systems theory, time series prediction, and theoretical biology.  At the moment I am mainly working on projects in two areas:  One is in financial economics, where I am developing a non-equilibrium theory for financial markets, and the other newer area is in the evolution of technology.
 
My projects in financial economics draw on ideas and methods from physics and biology as well as economics.  Financial trading strategies undergo descent with variation and selection, and provide a quantitative arena in which to study social evolution, a phenomena that is on one hand very different from biological evolution, but on the other, has many features in common.  My interest in financial markets is driven significantly by the availability of good data sets, which make them a good laboratory to study human decision making.  My group is studying several high frequency data sets from the London, New York, and Taiwan Stock Exchanges.  To varying degrees these data sets contain information about the identity of the participants, which makes it possible to track their behavior through time.  From a more applied point of view I think that a non-equilibrium theory for financial markets is essential if we are ever to properly understand phenomena such as volatility and risk.  My current research includes several projects.  For example, I am working on constructing a quantitative theory for market impact (price elasticity), which can be viewed as the interaction rule for agents in financial markets.  Another project studies the evolution of mutual fund size and its role in generating the distribution of trading volume.  In yet another we are trying to classify financial strategies and study their evolution over a ten year period in the Taiwan Stock Exchange.  This research is supported by an NSF grant in the interdisciplinary Human and Social Dynamics program and by Barclays Bank.  
 
More recently I have become interested in technology progress functions, a phenomenon that spans economics, engineering and complex systems.  Factors relating to the performance of a given technology tend to improve as a function of the number of units produced, typically following trajectories that are reasonably well approximated as power laws.  However, some technologies, e.g. those associated with computer information processing, appear to follow exponential improvement curves.  Here I am using “technology” broadly, to mean a variety of different things, ranging from the cost efficiency of a solar cell to computer performance to the output of an assembly line process to the productivity of someone operating a lathe.  There is no good theory explaining why technology performance curves should improve in the way that they do, and what factors influence the observed variations in rates of improvement.  To try to answer these questions I am pursuing a research program consisting of three parts, including empirical analysis of historical data, construction of theoretical models for technological evolution, and construction of portfolios in the presence of nonlinear feedback phenomena such as lock-in.   An area of particular interest is energy technology, and in particular improving public and private R&D strategies for reducing carbon emissions.  This research is supported by an NSF grant under their program on the “Science of Science and Technology.”