Scale-free avalanche dynamics in the stock market

Self-organized criticality (SOC) has been claimed to play an important role in many natural and social systems. In the present work we empirically investigate the relevance of this theory to stock-market dynamics. Avalanches in stock-market indices are identified using a multi-scale wavelet-filtering analysis designed to remove Gaussian noise from the index. Here, new methods are developed to identify the optimal filtering parameters which maximize the noise removal. The filtered time series is reconstructed and compared with the original time series. A statistical analysis of both high-frequency Nasdaq E-mini Futures and daily Dow Jones data is performed. The results of this new analysis confirm earlier results revealing a robust power-law behaviour in the probability distribution function of the sizes, duration and laminar times between avalanches. This power-law behaviour holds the potential to be established as a stylized fact of stock market indices in general. While the memory process, implied by the power-law distribution of the laminar times, is not consistent with classical models for SOC, we note that a power-law distribution of the laminar times cannot be used to rule out self-organized critical behaviour.