Systemic risk transmission : visualizing vulnerability

In just over 20 years, due to heightened globalization, the world economies have experienced over 30 widespread crises, stunting growth and employment, among other things. The many financial innovations that followed each passing cycle of a crisis, brought newer means of facilitating higher systemic risk, and the potential for newer contagion to be mired deep into the system. This poses an ever-growing challenge for investigators while naturally triggering higher research interests in this domain. This thesis, addresses some long-standing questions stemming from the systemic risk and contagion literature. The thesis presents a critical review of the relevant systemic crisis literature (Chapter 2) and focuses on the limitations and gaps in the extant literature. The following three concerning issues are identified: 1. Does the emerging picture during a crisis breakout show a common pattern to past episodes? 2. Can we disentangle sources of potential crisis from the intricately complex web of connections across international equity markets? 3. We progress to examine whether investors' risk preference induces a crisis and to what extent such predictors may indicate a pandemic? We progress gradually by addressing each of the above mentioned issues and proposing means for regulators and managers of risk to contain risk well before a crisis erupts. The first of the thesis essay (Chapter 3), develops a means of visualising the vulnerability of complex systems of financial interactions globally, using neural network clustering techniques. We aim to investigate 'if the emerging picture during a crisis breakout shows a common pattern to past episodes?' We show how time-varying spillover indices can be translated into two-dimensional crisis maps. These crisis maps have the advantage of depicting the changing paths of vulnerability, including the direction and extent of the effects between source and affected markets. We develop these crisis maps using equity market data for 31 global markets over the period 1998-2017. In this chapter, our aim is to convincingly implement means by which managers of systemic risk can simulate the effects of alternative intervention paths in a network and have some knowledge of where the most effective interventions may lie. The second essay of this thesis (Chapter 4) differentiates between 'good' and 'bad' interconnectedness by showing how signed spillover measures capture additional information compared to the unsigned spillovers. This builds the framework to address our second concern as 'we disentangle sources of potential crisis from the intricately complex web of connections across the international equity markets.' We analyze the behaviour of 30 global equity markets and compute multiple spillover measures using daily data over the period 1998-2017, which encapsulates many large and small crises. We use the signed realised volatility estimates to distinguish the contagious markets from the interdependent markets. Instead of relying on ex‚Äö-postcrisis information, we allow our model to identify crisis periods. It is clear the model efficiently detects crisis and newly emerging contagion in the system. The third essay of this thesis (Chapter 5) develops a means of visualising the vulnerability of complex systems of financial interactions, resulting from the changing risk tolerance of investors. As such, the investors' risk behavior contributes in the buildup of vulnerability in crisis and in calm periods. In this chapter, we examine if investors' risk preference induces a crisis and if yes, to what extent such predictors may indicate a pandemic ?' We show how both time-varying risk tolerance and spillover indices can be translated into two-dimensional information transmission and crisis transmission maps, respectively. Taken together, the information transmission maps have the advantage of proposing predictions to potential crisis transmission pathways in the crisis transmission maps. These maps provide easily digested visualization showing how information transmission predates crisis transmission, drawing from conditional signed spillover and risk tolerance indices computed from the equity market data for 31 global markets spanning from 1998 to 2017. Brought together, these approaches may help policy-makers and intermediaries take appropriate steps to subdue a crisis before it emerges.