THE EFFECTS OF HETEROGENEOUS BANKING NETWORK STRUCTURE ON FINANCIAL STABILITY: AN AGENT BASED APPROACH

• Main Author: Natali Susanto, Erwin
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/36616
• Institution: Institut Teknologi Bandung

Financial stability is the potential problem from the interconnected financial institutions on banking network system for central banks and regulators. To maintain the financial stability, central banks and regulators need to do action to minimize the impact of financial shocks transmission through contagion effects. This study propose a heterogenous banking network structure for testing the banking system stability. The heterogeneous banking system viewed as more realistic approach to current banking system compared to homogeneous banking system. For modeling the heterogeneous banking system, this study use core-periphery model (CP). This study view the response of banking system to idiosyncratic and systemic shocks (stress testing method) by using agent-based modeling in the simulation. In this study, each bank (agent) is created as an agent with learning, which can adjust the action by using reinforcement learning method. The results from the simulation in this study can give deeper insight of heterogeneous banking system’s response in shock for either on individual bank, central banks or regulator’s perspective. This study is a continuation of the previous study, which is conducted by Raswan and Koesrindartoto (2016). In previous study, they develop a dynamic stress testing banking model using the homogeneous banking network. The model follows the basic model that was developed by Novanto and Koesrindartoto (2012), which includes the effects of interbank lending and borrowing. However, the model in previous study is not suitable for modeling the real banking network. So, this study will develop a new model that is suitable for modeling the real banking network and then this study will explore the effects of that model. This study found that: first, mostly, lower connection probability between banks is better for maintaining financial stability in the simulation under idiosyncratic shock. Second, mostly, higher connection probability between banks is better for maintaining financial stability in the simulation under systemic shock. Third, we suggest that the regulator should control the range number of connection probability between banks by increasing the connection probability between small banks and decreasing the connection probability between large banks, connection probability between large bank and small bank in order to minimize shock transmission in the banking system in Indonesia.