Unveiling the connectivity structure of financial networks via high-frequency analysis

The paper deals with the problem of reconstructing the internal link structure of a network of agents subject to mutual dependencies. We show that standard multivariate approaches based on a correlation analysis are not well suited to detect mutual influences and dependencies, especially in the presence of delayed or propagative relations and when the sampling rate is sufficiently high to capture them. In particular, we develop and apply a metric based on the coherence function to take into account these dynamical phenomena. The effectiveness of the proposed approach is illustrated through numerical examples and through the analysis of a real complex networked system, i.e. a set of 100 high volume stocks of the New York Stock Exchange, observed during March 2008 and sampled at high frequency.