Problem On Stability Of Banks System

In July2007, with a series of collapses in financial sector, The Subprime MortgageCrisis hits the U.S. The crisis has caused a huge impact on American and even thewhole world’ s banking system as well as financial system. Most scholars attribute thiscrisis to excessive assets securitization, excessive credit creation in derivative products,and highly leveraged operational strategies adopted by ambitious investment bankers.These are indeed very important reasons, however, the crisis revealed the vulnerabilityand the unreasonable structure of banking system itself which is well worth exploring.This paper focuses on the banking system, and try to find the factors that lead to thefragility of the entire system from the internal structure of the banking system, and thustruly improves the structural problems of the banking system in order to reduce theprobability of the systemic banking crisis’ reappearance.On the other hand, previous studies mostly regard external shocks, such as macroe-conomic downturn and assets depreciation, as direct incentives of liquidity shortage inbanking system; also, most studies believe that the spread of crisis will finally end in alack of liquidity. However, when there is neither drastic macroeconomic volatility, norsignificant shocks from external environment, whether liquidity risk will also accumu-lates gradually? If such risk truly exists, what factors may affect the formation of suchrisk and how to control the accumulation of the risk as well as reduce it?In order to answer these questions, this paper constructed a heterogeneity bankingnetwork which breaks through the limitation that the homogeneity network does notaccord with the reality and adopted complex network models and agent-based modelsto analyze the features of the banking system. Firstly, this paper constructed an artifi-cial banking network，which is closer to real banking system based on the analysis ofinternal bank network structure and modeled on the mechanism of risk-proliferation.Then, more factors that influence bank’s financial status parameters in the process ofcrises diffusion are taken into consideration in the model and a more realistic mecha-nism of loss spreading can be designed. When it comes to the formation of liquidity risk in inter-bank payment system, with the help of agent-based modeling, all the director indirect participants involved in the payment system can be abstracted as differentagents in the system, and through differentiated properties and behavior preferences ofeach individual bank customer, better simulations can be carried out about the differ-ences between every bank’s business strategy and the differences between customers’behaviors. In this way, more useful and reliable results can be obtained compared withthe homogeneous model.Through analysis of the inter-bank network structure, the results showed that thereal banking network does have a characteristic of high clustering. Some banks are ata relatively central location in the network, and once they are in trouble, the crisis willcause tremendous losses to the entire banking system. However, it is not enough torecognize the important banks in the system solely based on “too big”. The size of abank only reflects part of the information; to analyze the systemic importance of a bankstill requests participatory analysis of the interaction between banks. Thus the saying“Too participated to Fail” should be more reliable than “Too big to Fail”.By doing impact tests for heterogeneity banking network, we found that the sys-temic loss of the banking system in crisis diffusion process is heavily affected by themicro financial structures of the banks that make up the banking system. The net assetlevel of banks have great influence on the total loss of the entire banking system afterthe crisis, and the overall systemic loss can be decreased by raising banks net asset lev-el. Different from previous findings, the results in this paper show that the relationshipbetween total loss and interbank exposures is not monotonic increase or decrease. Inthe period of stable macroeconomic environment, it is acceptable to encourage banksto carry out more lending and borrow business; in period of risky macroeconomic en-vironment, we should limit the proportion of inter-bank exposures, to prevent the rapidproliferation and spread of crisis and thus avoid huge losses in the economic system.Moreover, the relationship between banking network connectivity and systemic lossis also not monotonic increase or decrease. When the network connectivity level islow, increasing the total number of edges in the network will raise the chance of bankdefaults contagion. However, when network connectivity is already at a high level, fur-ther increasing the connectivity of the network will strengthen the digestion of external shocks in the banking network.We also found that without complementary external liquidity market, the networkconnection formed by inter-bank lending and borrowing can be crisis spread channelsnot only when the system is affected by external shocks. Even when there are noexternal shocks, liquidity risk in the inter-bank payment system can also be graduallyaccumulated due to the uncertainty of customers’ demand to pay. To eliminate thisrisk, we can make use of external liquidity market to adjust the inter-bank liquidity.The introduction of external liquidity market has a significant effect on eliminating thecongestion in inter-bank payment system; but this can be true only if the average costsof delayed payment are relatively higher than that of external financing. In addition,a stable external liquidity market can guarantee the stability of the inter-bank paymentsystem. Thus, constructing an efficient and stable inter-bank lending and borrowingmarket has profound and practical significance for maintaining the stability of China’sinter-bank payment system.