| Since the Mexican financial crisis in 1995 to the US subprime mortgage crisis and the European debt crisis in 2010, all of the financial crises have shown that systemic risk spillover between financial markets have result in huge damage of the whole financial system. And at the same time it has a tremendous impact on the real economy. Continuous financial crises outbreak have drawn global attention to systemic risk. As the lack of preventive measures against systemic risk crisis exposed, the Basel committee promulgated Basel ?, in which preventing and mitigating systemic risk function as the fundamental objective of macro-prudential supervision.Research on systemic risk spillover effect among various financial system sub-sectors is conducive to monitor and trackthe systemic risk in the early formation of the financial risk. It is of great reference for macro-prudential regulation at the same time.From the research point of view, the researchon systemic risk spillover can be divided into two categories. One is based on the theory of network analysis to study the systemic risk contagion, and the other is based on the CoVaR model to study the cross-market spillover effect. Based on the minimum relative entropy method to estimate the exposure situation between different banks with their balance sheet, Upper and Worms (2004) use the network analysis method to simulate the process of the risk contagion among the various banks. By constructing CoVaR model, Chen Jianqing etal. (2015) make an research on the spillover effects among Chinese financial sub-industry. There are two problems about the network analysis,the first is the unavailability of the balance sheet data among different banks besides simulation. The second is the method mainly used to analyze the inter-bank spillover effects. Study of risk spillover effect with CoVaR model, although it can be set to capture the spillover effects between financial institutions, the quantile of CoVaR model is mainly used to describe the yiled distribution of the financial institution, as a result, thespillover coefficients set by theCoVaR model will be consistent no matter under what economic conditions.This article will analyze the spillover effects among bank, security, insurance and other financial institution such as leasing and trust with SDSEVaR model.As this model refers to the used the financial industry index daily closing price, the market data fully reflect the risk in financial markets as well as avoid the data unavailability of the network analysis.It is of great reference significance to the establishment of systemic risk early warning mechanism, the SDSEVaR model study the sizeand duration of systemic risk spillover effect between financial sub-sectors under different economic conditions.With the combination between theoretical analysis and empirical study, this article make an research on the financial systemic risk spillover effects. Firstly, based on industry data of SW, this articleconstruct the GARCH model to calculate the EVaR, namely VaR based on the Expectile model. This article represents the financial institution systemic risk with EVaR, which funciton as input variables to build SDSEVaR model. Then Select different quantiles to represent different economic conditions, the SDSEVaR model can be builtthrough two-stage quantile regression. With this model, it is possible to analyze the size and direction of spillover effect among various sub-sectors of the financial system. The systemic risk can be divided into two categories considering the source, the first is the impact of macro shocks such as GDP, interest rates on all financial institutions; the other is the systemic risk originated from a local financial institution. Shocked by one event, the institution spread the risk to the relevant financial institutions through various channels,result in the whole financial systemic risk. When building SDSEVaR model, this article will use the real estate index data and commodity index data of SW as the control variable. It is possible to analyze the spillover effect among various financial sectors without the influence shocked by the same macro event. Based On the above research, this article will introduce impulse function (IRFS) to the SDSEVaR model. The specific method refers to the study of Zeno Adams et al. (2014), therefore dynamic SDSEVaR model can be set to analyze the continuous spillover effects among the financial sub-sectors.This article is divided into five chapters, each chapter is organized as follows.The first chapter is an introduction, which describes not only the background and significance of the topic, but also includes the framework, research methods and innovation. The second chapter istheliterature review. This article makes the review mainly from two aspects of the relevant research. The first part is the definition, characteristics, transmission mechanism and themeasure method of systemic risk. The second part refers to recent research on the spillover effects of systemic risk.The third chapter is the design research. This section presents the empirical scheme, and models involved. This article use EVaR to represent the systemic risk of each financial institution. The EVaR can be calculated by the GARCH model. Based on this, Each EVaR of financial sub-sector will function as the explanatory variables and the explanatory variables to construct a static model SDSEVaR model. The spillovers direction and size can be drawn from the model aboved. Finally, the impulse response function (IRFS) will be introduced to construct the dynamic SDSEVaR model, which can be used to analyze the duration of therisk spillovers effect.The fourth chapter is empirical research. Empirical research will firstly analyze the descriptive statistics and financial time series correlation coefficient matrix. Next construct GARCH model to calculate the EVaR of financial sub-sectors. Then EVaR of financial sub-sectors will function as input variables, theEVaR of real estate index and commodity index will function as control variables,thestatic SDSEVaR model can be drawn by the two-stage quantile regression.The impact of common shocks won't affect the analysis of the direction and magnitude of spillovers. The impulse response function will be included to construct dynamic SDSEVaR model. The conclusion of duration of spillover effectcan be drawn based the dynamic model. Finally, to do a summary of this article, chapter five is about conclusion and outlook. Firstly draw the conclusinfrom the empirical findings, Secondly make policy recommendations on this basis, and given the present study shortcomings and areas for improvement. Lastly give the directions for future expansion.Thesis conclusions include the following four aspects:(1) in the analysis of spillover effect among financial sub-sectors, the banking sector performs the most significant, when subjected to the external shock,risk spillover of Banking is largest financial sub-sector, the order of the risk spillover are insurance, security, and multivariate financial (2) Banking, as an insurance company's distribution channel, the spillover effects of insurance companies is very significant compared to that of bank. The cooperation model between banking and insurance companies determines the banking performs no obvious advantage for the spillover effect. Spillover effect caused by the insurance industry gradually weakened, the spillover effect has been reduced by 50% 20 days later. (3) the security industry when subjected to external shocks, the order of the risk spillover are insurance industry, diversified financial, banking. In the analysis of spillover effects caused by multivariate financial, the securities industry is the most obvious influenced. Mainly due to multivariate financial index includes trust, leasing and other financial institutions, the relationship between these institutions and the securityseems more closely. (4) when subjected to external shocks, the order of the risk spillover are securities, banking and insurance industries, multivariate financial are mainly affected by the banking and securities industry induration process of spillovers, and 20 days later, thespillover effect has reduced by 50%.Thesis innovation mainly concludes thefollowing three aspects:(1) EVaR is selected as the systemic risk of each financial institution. Kuan et al. (2009) study has shown that EVaR (Expectile-based Value at risk) inflects more comprehensive, compared with the Extreme Tail VaR. (2) two-stage quantile regression method to used to study the systemic risk spillover effects between financial sub-sectors. Relative to the most used method CoVaR, the proposed two-stage quantile regression method can set different quantile to represent different economic situations. When building SDSEVaR, this model will include the index data of n the real estate and commodity index as control variable during two stages quantile regression process, result in the exclusion of common shock influence during the spillover effects analysis. (3) In this paper, the impulse response function (IRFS) is introduced to the SDSEVaR model, it can be used to analyze the degree of continued spillover effect between financial sub-sectors and the length of the duration, which is of great significant reference for the establishment of early warning mechanism.Paper shortage mainly includes the following two points:(1) Thisarticle proposes an indirect method to analyze the systemic risk spillover effect between various financial sub-sectors. As thedirect correlation between financial sub-sectors is their leverage, liquidity and mutual ratio and other assets and liabilities. The direct mechanism can be reflected by these finance data.However, these data of the financial institutions are not available every day, thus it is impossible to analyze the spillover effects with data about direct relationship between various financial sub-sectors. (2) In addition tothe involved financial sub-sector banking, securities and insurance, this article uses the multivariate financial index to replace the other financial sub-sector.With the rapid development of derivatives market recently,the financial derivative market play an important role in our financial market. However, index data about derivative is not involved, the main reason is that the derivative market developed later than the financial sub-sectors involved in this article.This article could be improved from the following two aspects:(1) the SDSEVaR model could combined with the financial data including the leverage ratio, liquidity and asset-liability ratio to comprehensively analyze the spillover effects among various financial sub-sectors (2) in addition to thebanking, security, insurance and multivariate financial involved the trust industry and leasing industry, the funds and futures and other sub-sectors could be included in the research to analyze the spillover effects among the whole financial system.. |
PDF Full Text Request