Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.654534
Title: Monitoring systemic risk and contagion in financial networks
Author: Shaghaghi, Ali Rais
ISNI:       0000 0004 5358 8290
Awarding Body: University of Essex
Current Institution: University of Essex
Date of Award: 2014
Availability of Full Text:
Access through EThOS:
Abstract:
The 2007 financial crisis has shown that economists have been behind the curve in regard to mapping, modelling and monitoring the highly interconnected and global financial system. The failure of financial institutions has led to fears of system failure from domino effects of one failed entity bringing down others. This has given rise to concepts such as financial contagion and "too interconnected to fail". The latter was cited when there were tax payer bailouts of large financial intermediaries (FIs). This thesis has adopted network models to analyse the structure and stability of financial markets. Systemic risk from financial contagion is analysed on the basis of network stability, measured by the maximum eigenvalue of the network of bilateral financial obligations between FIs, and the corresponding eigenvector centrality of the FI is a measure of its contribution to network instability. A Pigou type negative externalities tax is designed on the basis of the eigenvector centrality (EVC) for FIs to internalize the costs they inflict on others by their failure. Chapter 3 reconstructs the financial network for the credit default swaps (CDS) which have been implicated in the recent crisis. Both the financial contagion characteristics of the CDS network and the use of the EVC tax is studied. Chapter 4 investigates the role of large complex financial intermediaries (LCFIs) in their global operations in OTC financial derivatives markets. The bilateral flows are empirically calibrated to reflect data based constraints. This produces a tiered network with a distinct highly clustered central core of 16 LCFIs. In Chapter 5 a novel genetic algorithm optimisation method is implemented to empirically fit networks that satisfy multiple constraints arising from balance sheet data. As FIs belong to multiple markets, each of which has its own network topology, the aggregated single network is often not a true picture of the complex interconnections. Chapter 6 uses a multi-layer network approach to generalize the eigenvector centrality measure from a single network case.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.654534  DOI: Not available
Share: