This project aims to analyze the impact of medium and small banks interaction on a set of banks which is considered as the core of the network.
Empirical studies have analyzed how liquidity risks faced by individual institutions turn into systemic risk. Financial regulation focuses on the most important and “systemic” banks in the global network. However, to quantify the expected loss associated with liquidity risk, it is worth to analyze sensitivity to this channel for the various elements of the global bank network. A small bank is not considered as potentially systemic; however the interaction of small banks all together can become a systemic element.
This paper analyzes the impact of medium and small banks interaction on a set of banks which is considered as the core of the network. The proposed method uses the structure of agent-based model in a two-class environment. Data from the actual balance sheets of 21 systemic banks (such as BNP Paribas or Barclays) are collected and combined with 579 fictitious smaller banks.
A simulation of 12 three month periods representing a midterm time interval three years is projected. In each period, there is a set of behavioral descriptions: repayment of matured loans, liquidation of deposits, income from securities, collection of new deposits, new demands of credit, and securities sale. The last two actions are part of refunding process developed in this paper.
To strengthen reliability of proposed model, random parameter dynamics are managed with stochastic equations as rates the variations of which are generated by Vasicek model. The Central Bank is considered as the lender of last resort which allows banks to borrow at REPO rate and some ejection conditions of banks from the system are introduced.
Three different liquidity crisis scenarios (asset variation, forced illiquidity in specific banks, and trust crisis) have been simulated and the loss impact on bank classes is analyzed through aggregate values representing the aggregate of loans and/or the aggregate of borrowing between classes. Our results show that the three groups of European interbank network have not the same response, and that intermediate banks are the most sensitive to liquidity risk.