An investigation into systemic risks and its mitigation based on complex network and agent-based modelling

Significant events, such as crises created by breakdowns of financial systems, natural hazards, failures of the infrastructure systems, production shortfalls within supply chains create enormous amount of losses to affected communities, whose impact is not only confined to a single country or economic sector but more. Their statistics on social and economic damages are sobering and urge us to a better understanding of the systems we live in, identifying the sources and drivers of systemic risks, and finding measures to mitigate crises. To make a contribution toward this goal we present two projects using agent-based models. In the first project, we simulate a dynamical supply chain network and investigate whether diversification within the supply chain can help mid-level firms survive prolonged financial crises. To do so, we analyze the lifetime distribution of undiversified and diversified firms and compare the average lifetimes and the rates at which the midsections and tails of the cumulative lifetime distribution decay. We have found that the main benefit of diversification does not lie in increasing profits or lower operating costs, but helps mid-level firms more effectively manage the systemic risks they are exposed to, and help them survive longer in a competitive and uncertain business world. In the second project we are particularly interested to investigate the interbank market in Singapore. To do that we develop a numerical scheme that allows us to reconstruct the Singaporean interbank network structure based on day-to-day data of interbank borrowing and lending recorded by the Monetary Authorities of Singapore. We then simulate an interbank network, where banks are connected via interbank loans and invest into individually optimized portfolios comprising of one risky and one risk-free asset. This allows us to investigate the systemic risk introduced by leverage and interbank loans. To understand the effectiveness of regulations on bank liquidity and leverage introduced by Basel III, we compare the lifetime of bank for three different regulation scenarios: (1) unregulated leverage (2) penalized leverage and (3) forbidden leverage. We observe that, although banks in our model have complete market information they do not perform well due to highly leveraged investments. In addition we investigate the consequences of incomplete or wrong market information. We identify that the over- and underestimation of probabilities for successful risky investments has similar severe consequences with a sharp drop in lifetimes. Our results show that a maximum in lifetimes is reached when regulations completely forbid leverage. Additionally we find that interbank lendings have a stabilizing effect on the network. This can be concluded since the reduction of linkages between banks reduces the lifetime of banks. On the other hand, if the network size and therefore the number of linkages per bank are increased we observe increased interbank lending accompanied by longer lifetimes.DOCTOR OF PHILOSOPHY (SPMS