||In this study we construct return correlation network for cross market listed stocks, commodities and foreign exchanges. Unlike conventional studies of market contagion, this study attempt to analyze the information flow among stocks using social network topologies. The change in return correlation network can be manifested by network parameters such as eigenvector centrality, degree, betweenness, closeness, modularity and density. These parameters have interesting economic connotation in terms of information asymmetry, herding, and information flow, and they can offer far more rich observation than traditional VAR analysis in market contagion studies. For example, we may be able to answer question such as which stock is the most important stock in information transmission across-markets or across industries, or how the correlation structure changes before and after certain events. More importantly, we can also identify the relationship between network topologies and global markets’ return distribution.|
This is the first study to apply network analysis to cross market securities and cross products. The purposes are threefold:
1. To estimate the ‘importance’ of each stock in cross-market information transmission.
2. To observe changes in topologies at important market events using dynamic data visualization, and investigate the underlying causes.
3. To empirically test the relationship between return topologies and return attributes.