Latest Publictions
2014 May/20
Analyzing Interconnectedness in Korean Financial Industry Using Granger Causality Network
Research Papers 14-04
PDF
- Research Topic Financial Services Industry Capital Markets
- Page 110
Suh, Hyunduk
- Other Publications
- Summary
- Interconnectedness among financial institutions can amplify the effects of shocks in the financial market. Global financial crisis during 2007-2009 showed how this channel can increase the systemic vulnerability of financial market. In this paper, I analyze interconnectedness in Korean financial industry using empirical methods.
Interconnectedness in financial sector can arise from several channels. In terms of funding resources, securities companies and specialized credit finance business companies have the highest ratio of inter-financial sector borrowing to total funding. Securities companies and asset management companies have been actively engaged in transactions in call loan market, a non-collateralized short-term money market. However, a concern for systemic risk led government to announce a policy to move these non-bank players to collateralized Repo market by 2015. Over-the-counter derivatives are also considered to be capable of creating potential systemic risk, due to their complex, opaque, and illiquid nature. In Korea, they are traded mainly by commercial banks and securities companies, but there is yet no case that these instruments caused any systemic crisis in financial market.
I also use Granger Causality Network method to study interconnectedness in Korean financial industry. A network is defined according to the existence of Granger causality between any two firms, and various network measures are used to analyze the degree and the shape of interconnectedness. Data used for estimation is the Expected Loss Ratio of financial institutions, and from sovereign CDS premium.
The main results from the Granger Causality Network analysis are summarized as the following. First, the degree of interconnectedness increases during the period of financial instability. Second, banks and financial investment companies mainly influence other financial sectors in the network, while insurance companies and savings banks are influenced by others. Third, the influence of foreign factors is low overall, but increased sharply during global financial crisis. Fourth, during savings bank crisis and Tongyang crisis, the influence of respective financial sectors increased. However, both events do not show severe degree of interconnectedness.
Granger Causality Network analysis relies on statistical concept of causality, thus it is susceptible to the criticism that it does not necessarily capture true causality. However, the approach has its own strength that it provides an intuitive, structural measure of interconnectedness in almost real-time. Considering these properties of Granger Causality Network, the method seems to be the most useful when used for an early warning signal indicator of financial instability.