|
Abstract:
|
Financial contagion is defined as a significant increase in cross-linkages after a shock to an individual financial entity. Networks, broadly understood as a collection of nodes and links between nodes, can be a useful representation of the financial system. In most of the related works in literature, network representation of the stock market is based on linear methods of association such as the cross-correlations of stock price fluctuations. We model a tail-based association network of financial linkages with contagion detected by using the method of Residual and Recurrence Times (RRT). Empirically, the resulting tail-based networks accurately capture the increase in U.S. market linkages at turbulent times, and discover new connections that linear methods of associations fail to detect.
|
