| Countries or regions have established complex relations through trade,credit investment and supply chain,etc.,which makes the world economic system highly interrelated and interdependent,and provides channels for the contagion of systemic financial risks.Once a financial market collapses or a financial institution collapses,or some mechanism or factor causes the contagion of systemic financial risk,it will cause huge losses to a country’s economy and even bring a global financial crisis.In the highly correlated global economic system,the network contagion of systemic financial risk makes the financial system of a country or region bear the impact of various financial shocks from all over the world at any time.With China’s deepening reform and opening up,China’s financial system has matured and established complex economic ties with the world economic system.On one hand,the complex relationship with the world economy makes China’s financial system face more uncertainty and more impact and pressure of external financial shocks.The risk contagion caused by external shocks can easily affect China’s financial system.On the other hand,with the continuous improvement and development of China’s financial system,banks,securities,insurance and other financial institutions infiltrate each other in business and cross integrate financial products,which makes it easy for cross infection of financial risks within China’s financial system.When the financial system encounters internal and external shocks,financial risks are likely to endanger the safety of the whole financial system through network infection.Therefore,it is of great theoretical and practical significance to study the spillover effect and dynamic contagion of systemic financial risk,and then provide policy support for preventing and resolving systemic financial risk.The spillover and contagion of systemic financial risk have the characteristics of complexity,dynamics,randomness and ripple effect.Therefore,a large number of literature at home and abroad use complex network technology to study systemic financial risks,such as asset liability network,Pearson correlation coefficient network,fuzzy financial network,tail risk network,spillover effect network,minimum spanning tree network,tree network and so on.From the existing literature,most studies research systemic financial risk by capturing the correlation,spillover effect and causality between financial institutions or financial markets,and there are still many research deficiencies.Firstly,the contagion of systemic financial risk has ripple effect,which is a dynamic process of gradual diffusion in the financial system from the source of contagion.Few literatures have studied the dynamic path of this process.Secondly,from the dynamic perspective,most of the studies in the existing literature use the rolling window method to study the differences of systemic financial risks in different periods.Few literature study how the characteristics of systemic financial risk change dynamically in the contagion ripple-spreading process.Thirdly,most of the complex networks constructed in the research are single-layer networks,and there is insufficient research on the spillover effect and dynamic contagion of systemic financial risk under diverse information.Although some studies try to make up for this deficiency by constructing multi-layer complex networks,they do not further extract relevant information based on the constructed multi-layer networks to study the multi-layer ripple effect of systemic financial risk.Finally,systemically important financial institutions play an important role in financial risk contagion and are also the key regulatory objects to prevent the occurrence of systemic financial risks.The existing literature on identifying systemically important financial institutions can be divided into two categories: one is to measure the contribution of financial institutions to systemic financial risk through spillover effect network;the other is to study the network centrality of financial institutions by constructing complex networks.The former only considers the direct spillover effect between financial institutions and ignores the indirect spillover effect;the latter lacks the research on analyzing the network centrality of financial institutions by constructing complex networks based on contagion ripple-spreading process.This paper proposes multiple complex networks to study the information spillover effect between Chinese financial institutions,the financial risk contagion under internal and external shocks,and the systemic importance of Chinese financial institutions.The multiplicity of networks includes two meanings.One is the diversity of network types,including random networks and non-random networks.Non-random network is a multi-layer information spillover effect network between financial institutions,which is used to study the information spillover effect between financial institutions;the random network is a contagion ripplespreading network model,which is proposed based on the multi-layer information spillover effect between financial institutions and considering other factors affecting financial contagion.It is used to study the ripple-spreading process of financial risk from the source of contagion to the whole financial system.Another meaning of network multiplicity refers to the multilayer of the network.Both information spillover effect network and risk ripple diffusion network are multi-layer networks.Although many literatures have studied the spillover and contagion of systemic financial risks by building multi-layer complex networks,such as multi-layer financial market networks,multi-layer interbank networks and multi-layer information spillover effect networks,these networks are non-random networks,and no further information was extracted to model and analyze the randomness and ripple effect of financial contagion.Based on the information spillover effect network and contagion ripple-spreading network of financial institutions,this paper analyzes the systemic importance of financial institutions from the perspective of spillover and network centrality.In the spillover analysis,this paper comprehensively considers the direct spillover effect and indirect spillover effect between financial institutions,and puts forward the collective spillover effect method;in the centrality analysis,a comprehensive centrality method based on heterogeneous contagion ripplespreading network is proposed.The research contents of this paper are as follows:Chapter 1 is the introduction,which explains the background,research significance,research content and technical route,as well as the innovation of this paper.Chapter 2 is literature review,which reviews the meaning and characteristics of systemic financial risk,discusses the literature on the measurement of systemic risk and financial contagion,as well as the identification methods of systemically important financial institutions.Finally,this paper summarizes the existing studies.Chapter 3 uses the stock historical data of financial institutions to build multi-layer information spillover networks,including return spillover layer,volatility spillover layer and tail risk spillover layer,to analyze the information spillover effect of financial institutions.Information spillover effect is obtained by constructing vector autoregressive model and then decomposing variance.Because the constructed vector autoregressive model contains many equations and needs to estimate many parameters,this paper introduces the LASSO-VAR to estimate the model.This chapter first analyzes the overall information spillover effect between financial institutions,then analyzes the net spillover effect from the perspective of asymmetry,and finally constructs a multi-layer information spillover network by threshold edge filtering method to study the differences between different information spillover effects.Although the multi-layer information spillover network constructed in this chapter is a non-random network,which cannot describe the randomness of financial contagion,through the research of this chapter,it can be concluded that which financial institutions receive the greater spillover effect from the financial system and which financial institutions are more vulnerable to financial contagion.And that is to lay the foundation for proposing contagion ripple-spreading network model.Chapter 4 is the study of contagion ripple-spreading network model.In view of the difficulty of modeling and analyzing the ripple effect and randomness of financial contagion in non-random networks,this chapter puts forward the contagion ripple-spreading network model,comprehensively considering the size of financial shock,the amplification effect of financial institutions on financial contagion,the ability of financial institutions to resist risk contagion,and the speed of financial institutions’ spreading financial risks,and the market distance between financial institutions.Then,this chapter studies the relevant characteristics of the model,and gives two ways of generating heterogeneous contagion ripple-spreading network for studying the differences between financial institutions in financial contagion and identifying SIFIs.Chapter 5 and Chapter 6 take Chinese financial institutions as the research object,and make an empirical study on the contagion ripple-spreading process under internal and external shocks.In Chapter 5,crude oil market is set as contagion source to study how the systemic risk contagion will spread in China’s financial system once the crude oil shock causes it.According to the multi-layer information spillover effect network,Chapter 5 successively studies the contagion ripple-spreading process based on return,volatility and tail risk spillover effect,and studies the differences of different contagion ripple-spreading process by constructing multilayer heterogeneous contagion ripple network.Chapter 6 studies the problem of how financial risks will gradually spread from the source of contagion in the financial system if an extreme crisis occurs in a financial institution and leads to financial contagion.Chapter 7 analyzes the systemic importance of financial institutions based on multi-layer information spillover effect network and contagion ripple-spreading network.This chapter studies the systemic importance of financial institutions from two aspects,one is from the perspective of Risk Spillover,and the other is from the perspective of node centrality of contagion ripple-spreading networks.In the Risk Spillover Analysis,considering there are direct spillover effects and indirect spillover effects between the financial institutions,this chapter proposes using collective spillover effect to identify systemically important financial institutions.Through the collective spillover effects,we can get the risk influence of financial institutions on the financial system and the risk pressure borne by financial institutions from the financial system.Financial institutions with greater risk influence have higher systemic importance.At the same time,financial institutions under greater risk pressure are also the focus of risk prevention.In the node centrality analysis,we first select the heterogeneous contagion ripple-spreading network through the method mentioned in Chapter 4,and then calculated which nodes have high centrality.To measure the node centrality from multiple dimensions,this chapter combines the centrality comprehensive evaluation proposed by Huang et al.(2018)with the heterogeneous contagion ripple-spreading network.Systemic importance of financial institutions are evaluated by calculating node centrality.Chapter 8 gives some policy suggestions on systemic financial risk prevention according to this study.Risk prevention policy recommendations include two parts: one is risk prevention policy recommendations from the perspective of dynamic financial contagion,and the other is some suggestions on the supervision of systemically important financial institutions.Chapter 9 summarizes the full text,discusses the shortcomings of this paper,and looks forward to further research in the future.The empirical results show that the information spillover effect between financial institutions in the same industry is higher than that across industries.For the cross-industry spillover effect,banks have obvious information spillover effect on insurance institutions and other financial institutions.Securities institutions have obvious information spillover effects on banks,insurance and other financial institutions,and the intensity is the strongest among the cross-industry spillover effects.The spillover effects of insurance institutions mainly point to banks and other financial institutions.The spillover effects of other financial institutions mainly point to banks and insurance institutions.The risk contagion caused by external shocks first spread to other financial institutions,and then to securities,banking and insurance institutions.Although the cross-industry spillover effect of other financial institutions is not so strong,they are very vulnerable to the external environment.Banking,securities and insurance financial institutions are highly related and easy to form an interconnected community network structure,while other financial institutions are not easy to closely connect with banking,insurance and securities institutions.The risk contagion caused by internal shock is first transmitted between financial institutions in the same industry as the contagion source,and then transmitted to financial institutions in other industries.Generally speaking,financial institutions in the securities industry have the highest systemic importance,and other financial institutions except banks,securities and insurance institutions have the lowest systemic importance.The banks’ systemic importance evaluated by collective spillover effects is different from that by centrality comprehensive evaluation.In the latter,banks are more systemically important.This is because the risk cross contagion between financial institutions first occurs among banks and then spreads to insurance,securities and other financial institutions.The nodes representing banks have established more directed links,which means that banks are easier to spread financial risks.The innovations of this paper: firstly,in terms of method innovation,this paper proposes a multiple complex network model to analyze the spillover and contagion effects of systemic financial risk.The multiplicity innovation of complex networks is reflected in two aspects: on the one hand,the constructed financial network is composed of non random networks and random networks,on the other hand,it is reflected in the multi-layer nature of the network.The non random network is a multi-layer information spillover effect network,which is used to study the multi-dimensional systemic financial risk spillover relationship between financial institutions;stochastic network is a contagion ripple-spreading network model,which is used to study the ripple-spreading processes of systemic financial risk.Secondly,the contagion ripple-spreading network provides a new perspective for the study of financial risk contagion.The contagion ripple-spreading network model comprehensively considers the important factors affecting the transmission of financial risk,such as the size of financial impact,the amplification effect of financial institutions on risk,the speed of financial contagion,the ability of financial institutions to resist financial risks,and the market distance between financial institutions.It can warn the transmission path of various potential contagioin sources.The contagion ripple-spreading network model can intuitively give the dynamic path of financial risk from the source of contagion to the financial system in the form of dynamic network,which provides a new perspective for the study of syetmic risk.Thirdly,the contagion ripple-spreading network model proposed in this paper is also a supplement to the study of financial contagion using simulation technology.Traditional simulation methods mainly study the possibility and severity of financial contagion in financial networks.Different from these studies,the contagion ripple-spreading network model proposed in this paper simulates and analyzes the ripple effect of financial contagion,and studies the contagion paths from contagion source to the whole financial system.Fourthly,from the perspective of empirical research,this paper studies the multi-dimensional information spillover effect between Chinese financial institutions from the perspective of multi-layer complex networks.Based on the proposed contagion ripplespreading network model,this paper studies how the contagion will spread gradually from contagion source to the whlole financial system,and the possible early warning paths if a potential contagion causes financial contagion.Fifthly,this paper proposes collective spilover effect method and heterogeneous network based centrality method to identify systemically important financial institutions.Collective spillover effect considers both the direct and indirect spillover between finanicla institutions.Centrality method integrate a variety of centrality measures to measure the systemic importance of financial institutions in heterogeneous contagion ripple-spreading networks. |
PDF Full Text Request