Interbank Contagion And Risk Immunization: A Quantitative Framework Based On Network Theory

Interbank market is an essential financial intermediate for financial system to function healthily and stably, and is also an important component of modern financial systems. In individual aspect, interbank market can regulate the liquidity surplus or short for individual bank, such as the banks with liquidity surplus can lend to the banks with liquidity short. In financial system aspect, capital supply and demand of financial market can be reflected through interbank market. In this way, it will provide central bank a platform to establish and implement monetary policy, such as the open market operations. However, at the time to exert its function, interbank market increased the relativity between the risk exposures of banks. Out of question, it will present banking system potential risks, and will provide paths for risk contagion in bank failure.At present, Chinese interbank market develops rapidly. How to avoid risk contagion in banking systems is always an important problem faced with Chinese financial regulators. On the one hand, China is promoting financial reform step by step, and foreign banks will also participate in Chinese financial market. In this way, there will be a more close relationship between foreign banks and the banks in the country. This will also provide new paths for bank risk contagion. On the other hand, the market scales and the members are all increasing in Chinese interbank market, and furthermore, the financial tools and the trade forms are more and more complex. Therefore, a complex relationship of lending and borrowing between bank agents will be formed in interbank market. This will put forward new challenges for bank regulation definitely. Consequently, with these new positions, it is important to study and understand the mechanisms of risk contagion in interbank market.However, most of the researches in banking risk contagion are focused on qualitative or empirical analysis till now. It is devoid of quantitative analysis in the research fields of bank contagion. In the view of the relationship between the structure and function of interbank market, this paper presents a quantitative framework for the analysis of banking risk contagion. Specially, the quantitative framework is put forward based on network theory. In this framework, we have studied the risk contagion mechanisms and immunization strategies systematically. The main work and conclusions are presented as following:1. This paper summarized the existed literatures on studies of bank risk contagion. Firstly, we reviewed existed researches on the features of interbank market structures. Secondly, this paper summarized empirical and theoretical studies on bank contagion in detail. Moreover, empirical methods on the research of bank contagion have been reviewed, and the shortages of the methods also presented. Finally, with the reviews in this paper, we point out that, the latest directions in the research of the issues on bank contagion are the studies based on network theories.2. A banking network model used to describe the structures of interbank market is presented in this dissertation. In this paper, the network theory is used to describe the structures of interbank market. In this way, most of the interbank network models have been wholly introduced to describe interbank market, such as random banking network model, small-world banking network model and scale-free banking network model, etc. Moreover, based on the existed empirical results, we put forward a two-power-law banking network model in this paper. In our research, we study the mechanism for the existence of the two-power-law phenomenon. Additionally, we explain the economic significance of this result is that, the closer of the interbank lending and borrowing ability between two banks, the easier they will establish a relationship of interbank lending and borrowing. Furthermore, comparative studies on micro-structures between banking network models have been presented in this paper. Our results reveal that, in the banking network models with different macro-structure features, there are great differences in types and amounts of micro-structures.3. A bank agent model based on balance-sheet connections is put forward in this paper. Given the distribution function of interbank credit liabilities, and with three parameters such as p,q and m, we can obtain the balance-sheet structures of all banks in interbank market. In this way, we put forward an agent model to describe the behavior of a bank in interbank market. Furthermore, based on this agent model, the analysis of bank contagion based-on balance-sheet connections has been given.4. A quantitative framework is presented through combing the Macro-Structure of interbank market and the Micro-Agent of banks. Based on the banking network model and the bank agent model, and with the thoughts of combining the Macro-Structure of interbank market and Micro-Agent behavior of banks, a quantitative framework has been presented for the analysis of banking contagion. That is the MS-MA framework. Furthermore, we put forward the quantitative analysis method for bank contagion under the MS-MA framework.5. We studied mechanisms of bank risk contagion base-on the MS-MA framework. Under the MS-MA framework, this paper studied how the banking network structure and bank agent behavior influences the mechanisms of bank risk contagion. Moreover, with different patterns of contagion, we analyzed the bank contagion process respectively based on default pattern and the pattern of initial shock apportion. Our results indicate that, in the long term, increasing in the member size of interbank market can help to lower the contagion effect in the market. Moreover, in the interbank market with large member size, bank agent both participating in interbank lending and borrowing can help to reduce risk exposures and decrease contagion effects more effectively, comparing with the bank agent only take part in interbank lending or borrowing. Furthermore, whether the decentralization measures can effectively lower the contagion effect relies on the states and structure features of the interbank market.6. The immunization strategies of bank risk contagion have been studied under the MS-MA framework in this paper. Under the quantitative framework, we have studied the target immunization strategy,the netting immunization strategy,the interbank bailout immunization strategy and the mix immunization strategy. With the research of the target strategy, it reveals that the banks with large size of bank assets and high degree of connections should be immunized firstly. Because these banks will produce huge shocks to banking network system when it failed, so it cannot be failed. This means Too Systemic To Fail. Moreover, comparing with no immunization strategies, taking the strategies such as the netting strategy,the interbank bailout strategy or the mix strategy can all obtain a better immunization effect. However, with the same pattern of contagion and the same structure features of interbank market, different immunization strategy will produce different immunization effect. With synthetic consideration, mix strategy can generally obtain a better immunization effect.The innovative points of this dissertation are as following:1. This paper put forward a two-power-law banking network model to describe the structure of interbank market. Empirical study reveals the existence of the two-power-law degree distribution in interbank market, but no theoretical model to explain this phenomenon. In our research, we put forward a network growth model and study the mechanism for the existence of the two-power-law phenomenon. Additionally, we explain the economic significance of this result is that, the closer of the interbank lending and borrowing ability between two banks, the easier they will establish a relationship of interbank lending and borrowing.2. A quantitative framework has been presented for the analysis of bank contagion. Presently, there is no quantitative framework for the analysis of banking risk contagion. In this paper, we systematically put forward a quantitative framework to analyze the problems of bank contagion, based on the thoughts of combining the Macro-Structure of the interbank market and Micro-Agent behavior of the banks. In detail, this framework is presented based on two models, that is the banking network model and the bank agent model.3. Based on the MS-MA framework, banking risk contagion and immunization strategies have been quantitatively studied in this paper. On the one hand, how the structures of interbank market and the behaviors of banks affect the banking risk contagion process has been studied quantitatively. And moreover, under the different patterns of risk contagion, we analyzed the bank contagion process respectively based on default pattern and the pattern of initial shock apportion. On the other hand, to control the banking risk contagion process, four immunization strategies such as the target immunization strategy,the netting immunization strategy,the interbank bailout immunization strategy and the mix immunization strategy have been put forward in this paper. Furthermore, immunization effects of the strategies have been studied under different structures of interbank market.