Abstract | After the financial crisis of 2008, the connectedness of financial institutions has become an academic concern. However, the traditional banking model is difficult to describe the connectedness of financial institutions, so network analysis has become an important method in the study of systemic risk. Technically, there are two challenges to precision regulation. One is that networks constructing with different data will express different information. The other is that the data constructing networks will be driven by common factors and make the results incorrect.
In order to meet the above challenges, we define two kinds of networks: market network based on market data and balance sheet network based on balance sheet data. We conduct an empirical study on 33 financial institutions listed in A-share stock market of China before December 31, 2010. Firstly, we refer to Hale & Lopez's (2019) method to decompose stock return and ROA by CAPM regression and MIDAS model respectively. Then we respectively use DY method (Demirer et al., 2018) and TMFG technology (Massara et al., 2016) to construct financial networks. Finally, we compare the graph, nodes’ systemic importance and the total connectedness in different networks.
This paper contributes to the literature in three ways. Firstly, we compare market networks and balance sheet networks under a unified framework. Using the basic method of network analysis, we analyze the roles of financial holding groups in different periods and the dynamic changes of the systemic importance of financial institutions with different sizes. Secondly, we use mixed data sampling (MIDAS) model in the data processing of systematic risk research in this paper. The application of MIDAS model improves the accuracy of relevant indicators about systematic risk. Thirdly, we introduce Triangular Maximally Filtered Graph(TMFG)into the study of systemic risk for the first time. TMFG has advantages over other technologies such as Minimum Spanning Tree, which greatly improves the properties of the model.
From the network analysis, we can draw the following conclusions: (1) While the market network emphasizes the role of shareholders behind the financial institutions in the network, the balance sheet network highlights the role of the size of financial institutions. (2) During the sample period, the systemic importance of financial institutions whose big shareholder is state-owned big banks or insurance companies is low, while the systemic importance of the financial institutions controlled by non-financial state-owned groups is high. Moreover, in the period of financial turbulence, the systemic importance of the financial institutions whose big shareholders is the private-controlled groups and the local government-controlled groups will significantly increase. (3) There isn’t a linear relationship between size and systemic risk. During periods of financial stability, only very large financial institutions have high eigenvector centrality. However, in the period of financial instability, the eigenvector centrality of small-scale financial institutions will increase significantly. (4) The systemic interconnectedness index expressed by the market network with a lag of one quarter generally follows the trend of that of the balance sheet network. What’s more, the systemic interconnectedness index is found to be relatively sensitive to the monetary policy of the federal reserve.
Based on these findings, we propose four policy recommendations. Firstly, in order to improve regulatory efficiency and make correct decisions, regulatory authorities should make full use of different information provided by different networks. Secondly, regulators should strengthen the supervision of financial holding groups, including setting special requirements on the leverage ratio, strictly managing the equity structure, and strengthening the supervision of related transactions of financial holding groups. Thirdly, regulators should strengthen supervision of the largest financial institutions in times of financial stability. However, in times of financial instability the supervision of small and medium-sized financial institutions should be strengthened. Fourthly, regulators should pay close attention to the impact of external shocks, especially the monetary policy of the federal reserve, on China's financial stability. Regulators also should prevent systemic financial crisis by the fed's monetary policy adjustment.
|