Asia Pacific University Library catalogue

A thesis submitted in fulfillment of the requirements of Asia Pacific University of Technology and Innovation for the award of Doctor of Philosophy in Finance (UCPF1810FIN).

The current study pursued a research agenda in view of improving the modelling of Early Warning System (EWS) for systemic banking crises by examining the causes for the failure of existing EWS models prior to the global financial crisis of 2008 (GFC 2008) as well as the novel findings in the aftermath of the GFC 2008. Using a sample of 18 Developed countries that were hit by the GFC 2008 and three control case countries of similar profile, the current study examined the robustness of EWS indicators considered in the aftermath of GFC 2008 and identified new EWS indicators through the application of three main methods of investigation, namely, the ‘Signal extraction’ approach, the Multivariate probit panel regression method, and the Binary Recursive Tree technique, supplemented by the application of a Vector Autoregressive model as well as statistical and balance sheet analysis. In the post-GFC 2008 era, several studies highlighted excessive credit growth as the primary source of vulnerability to the financial system. The current study suggested moving a step further in that what also mattered was where the credit flowed, both to and from.

In view of improving prediction of systemic banking crises, the proposed list of EWS indicators were integrated into a comprehensive EWS model for systemic banking crises that comprised of six main areas of surveillance. The proposed EWS model broadens the scope of surveillance from the narrow realm of key macroeconomic developments and the evolution of vital banking sector metrics to six main areas of interest that add the incidence of excessive growth in credit, asset price growth, expansion in shadow banking activities, and global liquidity, along with global exposure of banks, into the comprehensive EWS model.