Research on stock markets

Financial markets are doubtlessly one of the most complex structures in the known Universe. Apart from many properties that they share with other natural complex systems, they have an additional, unique property: their building elements - the investors - are intelligent beings. Therefore, the financial markets have exceptionally strong ability to self-organize. Any research aimed at describing and understanding them itself contributes to their self-organization via feedbacks. This poses a real challenge. On the other hand, the financial markets, and especially the stock markets, offer enormous amounts of clean data. Since they are designed and controlled by humans, there is also perfect knowledge of the rules they obey and the constraints imposed on their dynamics. These features make the financial markets be a particularly interesting subject of study.

Our activity in this field started with analysis of empirical, high-frequency data from different stock markets. We were interested in statistical properties of price fluctuations, correlation structure, and dynamics of correlations. We applied methods of statistical physics, multivariate analysis, fractal geometry, and graph theory. Having collected substantial experience in dealing with stock market data, now our interest naturally shifts to modelling it. We choose agent-based approach, which seems to be ideal for bottom-up modelling direction. Our ultimate goal is to create an artificial market that mimics all the main properties of real markets including the self-organization, while being as simple as possible.

List of our publications on stock markets can be seen here.