微观层面系统性金融风险指标的比较与适用性分析——基于中国金融系统的研究

准确测度金融机构对整体系统性金融风险的边际贡献是加强宏观审慎监管的基本前提。本文对常用的系统性金融风险指标进行了比较分析,并以“能否涵盖规模、高杠杆率和互联紧密性三方面信息”、“排序结果是否与银保监会认定的系统重要性银行名单相吻合”、“是否具有宏观经济活动预测力”三方面对上述指标在我国金融体系的适用性进行了综合评价。结果显示,SRISK更适于作为我国微观层面系统性金融风险的测度。同时,本文发现,“LRMES约等于1-exp(-18*MES)”的经验关系不具有普适性,不适用于我国金融体系。

Comparison and Applicability Analysis of Micro-level Systemic Risk Measures: A Study Based on China's Financial System

The 2008 global financial crisis had a significantly negative effect on the real economy, and the systemic risk in the financial sector attracted unprecedented attention from academics and policy makers. The high macro leverage and credit risk are currently prominent financial issues in China. Moreover, the interest rate hike and balance sheet reduction of the Federal Reserve and the U.S.-China Trade War are having negative spillover effects. As a result, China's regulatory bodies have begunimplementinga macro prudential policy to defend the bottom line of no systemic risk.
Although numerous studies have examined the systemic risk in China, the literature in this area has several shortcomings.First,although studies have used several risk indicators, such as MES, SES, ΔCoVaR, and SRISK, to quantify the marginal contribution that a single financial institution makes to the overall systemic risk, no studies have comprehensively evaluated the applicability of these indicators to China's financial system. Moreover, we find that there are prominent differences in the systemic importance rankings based on MES, ΔCoVaR, and SRISK. Second,studies have not back-tested the effectiveness of the above risk indicators from the perspective of negative externality, which is the most essential characteristic of systemic risk. Third, several studies simply use the empirical approximation “LRMES=1-exp (-18*MES)” proposed by Acharya et al. (2012) to quantify the SRISK of individual financial institutions in China. However, the approximation is based on the U.S. financial system and compatibility with China's financial system has not been seriously explored. Lastly, the literature uses either the capital adequacy ratio minimum requirement (8%) under Basel II or the historical average of prudent capital of various institutions to determine the proportion of prudent capital for all of China's financial institutions, which include commercial banks, security companies, and insurance companies, and has not carefully considered China's financial regulations.Therefore, we aim to fill these gaps by constructing micro-level systemic risk indicators that are applicableto China's financial system.
To solve the aforementioned problems, we address the following issues. First, we examine whether the empirical approximation “LRMES=1-exp (-18*MES)” is applicable to China's financial system,and if not, whether it is possible to derive a similar approximation for China.Second, we attempt to determine the proportion of prudent capital for China's financial institutions, and evaluate the applicability of the aforementioned systemic risk indicators to China's financial system.
First, using the derivation of the principle of the above approximation, we find that “LRMES=1-exp (-18*MES)” is not applicable to China for the following reasons: (1) the approximation, which is derived from the U.S.financial system, is not universal for all the economies; (2) if we define a systemic event as a stock market decline of 40% over 6 months, the approximation “LRMES=1-exp (-13*MES)” is applicable to China; (3) if we define a systemic event as a stock market decline of 10% over 1 month, the approximation “LRMES=1-exp (-3.5*MES)” is applicable to China.
Second, we determine the prudential capital ratios for banks, securities, insurance, and real estate companies as 11.5%, 18%,15%, and 20%, respectively. It is not reasonable to determine the same prudential capital ratio for all financial and real estate institutions because institutions in different sub-industries have different operating features and capital adequacy ratios.Thus, we need to determine specific prudential capital ratios for these institutions according to the relevant supervision requirements.
Last,and most importantly, based on the above findings, we conclude that SRISK is more effective than the other indicators in measuring marginal contributions that financial institutions make to the systemic risk in China for the following reasons.First, only SRISK can simultaneously cover information on the size, leverage, and interconnectedness of firms.Second,the top 20 SIFIs identified and sorted by SRISK are in line with the list of SIFIs identified by the CBIRC, while the results based on the other indicators are very different from the CBIRC list.Third, the back-testing results indicate that the “overall SRISK value” can effectively predict China's macroeconomic activities. Fourth, the results based on MES and ΔCoVaR suggest that the regulatory authorities should pay more attention to financial institutions with small market capitalization, large volatility, and strong interconnection when the market tail risk is rising, which obviously deviates from the prudential regulatory practice.