Fixed link
0
vote

Guest Commentary by Paul Ormerod, Author of 'Butterfly Economics'. Econophysics: Achievements, Dangers and Challenges

Paul Ormerod

posted on 06 September 2002

Econophysics has made a decisive advance in knowledge in terms of portfolio theory. Markowitz mean-variance analysis, based upon the assumption of normally distributed asset returns and upon the historical correlation matrix between assets, still dominates practical risk management in financial markets. But econophysics has blown it out of the water scientifically. Key contributions were made by the pioneers in this new field.

Econophysics is also giving us for the first time a sound agent-based approach for understanding why financial markets behave as they do: for example, minority game, the work at Santa Fe institute.


Dangers

One danger is that econophysics has already made its key advances in financial market analysis. But there is a large and rising number of papers which simply restate the themes which have already been established (e.g. many of the contributions to the Bali conference). The one truly outstanding challenge here, that of successful asset price prediction, comes into the category of 'too difficult': rather like trying to solve the question of unified field theory.

The second danger is that physicists will be seduced by the elegance of conventional economic theory, despite the fact that many of its assumptions are often rejected empirically. Rational expectations, utility maximisation, Nash equilibrium : all may seem plausible and enable neat theoretical models to be built. But economics itself is providing more and more empirical evidence that these assumptions lack empirical support (recent additions to a growing list are Mirowski, Machine Dreams, CUP, 2002, Tenorio and Cason, Economic Journal, 112, 2002, Frederick et.al., Journal of Economic Literature, XL, 2002). It would be the ultimate irony if physicists, whose own discipline has by far the firmest empirical foundations of all the sciences, were bamboozled into using refined assumptions which are often rejected empirically.


Challenges

Financial market analysis is only one very small sub-set of economics. The real challenge for econophysics is to tackle the much more substantial issues of economic theory : for example, growth, business cycles, the behaviour of firms and consumers. Examples do exist: Stanley and colleagues on the growth of firms, Zhang on the evolution of markets, Ormerod on the US business cycle. All these offer more powerful explanations of their respective problems than does mainstream economics.

Financial markets are comforting to physicists: large amounts of clean data, just as they are used to. The rest of economics isn't like that. Data series are usually short and messy. It's much harder to get good empirical confirmation of a theory, and just as hard to think of the theory in the first place. But it is really here that physics, with its insistence on firm empirical support, has potentially the most to offer.

Paul Ormerod, September 2002

pormerod@volterra.co.uk

www.volterra.co.uk