Seminar

Abstract:

Faced with the development of financial liberalization, the correlations between financial institutions are constantly rising. The subprime mortgage crisis of 2007-2009 shows that there exists high correlations between businesses of different banks. A failure of a specific bank might in turn cause another bank to fail, which could result in chain reaction. Under this background, it is of great significance to study the correlations in the financial system.

Related researches include building correlation indicators based on specific business associations between financial institutions or based on market data such as stock or CDS prices. Considering the financial market is a typical complex system, the above methods haven’t studied the topology of the network, so it can’t describe the financial market explicitly. Complex network theory provides another perspective, up till now many of which still focus on measuring the importance of single nodes of the system.

However, regulatory authorities gradually abandon supervision of the micro-risk of individual institutions statically. They are paying more and more attention to the interconnection and dependence of financial networks, which requires methods for network’s overall and dynamic analysis of systemic risks.

In order to satisfy this need, a percolation based financial network is constructed. This method can clearly observe the development process of network system. As percolation threshold increases, more and more institutions are connected in the network. Different threshold corresponds to different state of network. This research uses data of 19 financial institutions in China of 3,724 trading days, and analyzes both static and dynamic percolation thresholds. The results show that the value of percolation threshold consists with the operation of the financial system. Thus, percolation is a method for financial system analysis and percolation threshold can be used as an indicator of financial supervision.