基于银行-资产双边网络模型的系统性风险及投资策略研究

受近几年的国际金融危机及金融全球化的影响，对金融系统的系统性风险的研究已成为国内外学者的关注热点。考虑到当前基于"银行-资产-银行"间接传播渠道的相关研究相对匮乏，本文基于银行-资产双边网络模型来分析银行系统性风险。首先，使用中国47家上市银行2018年的资产负债表数据构建了中国银行系统的双边网络模型，研究分析各类资产遭受冲击时外部冲击、降价出售效应及银行所持有的各类资产占银行总资产的比例对银行系统性风险的影响。然后，引入系统性冲击方式，通过设置具有不同属性的两大类资产并生成四种冲击事件来构建银行的投资策略模型，从资产视角探讨银行最优的投资策略。研究发现，外部冲击与降价出售效应这两个产生系统性风险的影响因素在一定区间值时会产生叠加效应，使银行系统性风险急剧增加；五种资产类中，贷款类资产对外部冲击最敏感；分析发现在各类资产冲击下都未倒闭的所有银行的资产组合具有一定的相似性；进一步研究发现银行系统中存在着最优的资产组合，使得银行在稳定的同时能获取最大收益，并且资产负债比越大的银行其风险承受能力越强，从而可以选择更激进的投资策略来追求高收益。

Study on Systematic Risk and Investment Strategy Based on Bank-Asset Bipartite Network Model

Influenced by the international financial crisis and financial globalization in recent years, the research on the systemic risk of the financial system has become a hot topic. Existing related researches mainly focus on direct contagion channel of interbank market ("bank-to-bank"), while, the indirect contagion channel of "bank-asset-bank" is relatively little. Moreover, the research on systematic risk of Chinese banking system and quantitative investment strategy of banks based on the bilateral network has not been studied yet. Therefore, based on the bank-asset bipartite network model (there are two types of nodes:bank node and asset node, and the bank-asset bipartite network is constructed by the asset portfolio of each bank), the systemic risk and investment strategy of banks are analyzed in this paper. Firstly, the bank-asset bipartite network model of Chinese banking system is constructed by using 47 listed banks' balance sheet data in 2018, to study the systemic risk when each asset class is shocked. All five categories of banks (state-owned large-scale commercial banks, joint-stock commercial banks, urban commercial banks, rural commercial banks, and policy banks) except foreign banks in the Chinese banking market have been covered in the 47 listed Chinese banks used in this paper, and all listed banks in China are basically covered. The data comes from the balance sheet in the 2018 annual reports disclosed by each bank, including the total assets, total liabilities and detailed asset portfolio data of each bank. In this paper, the assets of each bank are finally divided into five categories based on their investment attributes and income types, namely, cash, interbank assets, financial investment, loans, and other asset classes. Then, the systemic shock is introduced, and the investment strategy model is constructed by setting up two classes of assets with different attributes and generating four shock events (in the investment strategy model, assets are divided into two categories:low-risk and low-return, high-risk and high-return; four kinds of systemic shocks are formed by the combination of "profit" or "loss" events of the two categories of assets). The impact of changes in the proportion of different assets invested by banks on the systemic risk of banks are studied to explore the banks' optimal investment strategy. Research results show that external shocks and price-cut selling effects, which are the two factors of the systemic risk, will produce a superposition effect at a certain interval value, causing a sharp increase in the systemic risk; the loan assets are found to be most sensitive to external shocks; the asset portfolios of all banks that have not failed under all of asset shocks have certain similarities; further research finds that there is an optimal asset portfolio in the banking system, which enables banks to obtain maximum returns while maintaining stability, and the greater the asset-liability ratio, the stronger their risk tolerance, and the more aggressive investment strategies can be chose to pursue high returns. The role of external shocks and price-cut selling effects is found, which provides a certain reference for the related research of the systemic risk under the bank-asset bilateral network. The research also shows that the bank's investment portfolio has a significant impact on the systemic risk of the bank, which is helpful to the bank's investment decisions and the risk monitoring of the relevant regulatory authorities. The actual data of Chinese listed banks are used for empirical analysis, which provides a reference for the Chinese banking industry to deal with external risks and asset price fluctuations.