Statistical Methods in Finance 2017

Dec 16 - 19, 2017


Understanding complexity of "market states"

by Anirban Chakraborti

"The understanding of financial markets as an example of a complex system has posed many challenges and interesting findings. We analyze financial data from the S&P 500 stocks and Nikkei 225 stocks for the period 1985-2017. We follow a methodology similar to the one proposed in the paper of Münnix et al. [1] and try to find the "state" for a financial market; specifically, to study changes in the correlation structure or identify "critical states". We find that there exist a number of characteristic correlation structure patterns in the observation time windows, and that these characteristic correlation structure patterns may be classified into several typical "market states" using similarity measures and multidimensional scaling maps. Furthermore, using the details of the multidimensional scaling maps and eigenvalue spectra, one can study the "complexity" of the market states.