Network-based approaches in modelling financial contagion

Wu, Fan ORCID: https://orcid.org/0009-0009-0002-7434 2026. Network-based approaches in modelling financial contagion. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF (Thesis) - Accepted Post-Print Version Download (17MB) | Preview
	PDF (Cardiff University Electronic Publication Form) - Supplemental Material Restricted to Repository staff only Download (1MB)

Abstract

The financial market is a complex system in which information, such as volatility risks and investor sentiment, propagates across assets. Understanding these transmission mechanisms and the resulting level of overall risk contagion is important for market participants and regulators in managing financial risks. This thesis uses network approaches in modelling financial market contagion and ultimately proposes a novel network-based sentiment and risk index. This research is motivated by the collapse of Silicon Valley Bank (SVB), which is treated as an exogenous shock that triggered the UStechnology sector. A comprehensive literature review is conducted to establish the theoretical and empirical foundations of this study. This thesis consists of three main empirical studies. We first apply the Diebold Yimaz network connectedness approach to examine the dynamic and static volatility risk contagion effect among the US technology sector. The properties of the contagion networks before and after the SVB failure are also analysed to assess the changes in the connectedness. Second, we propose a transfer-entropy-based network to accommodate the non Gaussian properties of the sentiment data and measure the information spillovers of news and social media across the selected technology companies. We analyse and compare the dynamic sentiment spillover networks over time, and the regime network analyses are conducted both across the different regimes within the same media source and between two different media sources. Motivated bytheeffectofthesentimentonthefinancialmarketandtheimplications of the market interconnectedness. We take the properties from the sentiment spillover networks and the volatility risk contagion networks to propose a novel network-based sentiment index and risk index. The proposed sentiment index demonstrates a stat istically significant leading effect on the market benchmark sentiment index, the Fear and Greed Index, while the proposed risk index shows a significant relationship with, ii and strong predictive power for, market risk measures, including the VXN and VaR95 GARCH. Fundamentally, this thesis demonstrates the significance of network modelling in capturing the complexity of the financial contagion phenomenon. The proposed framework complements existing approaches and introduces a new perspective for quantifying risk contagion mechanisms in financial markets, and it can be further extended to investigate sector-specific risk transmission processes

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Schools > Mathematics
Uncontrolled Keywords:	1. Network Analysis 2. Financial Contagion 3. Information Spillover 4. VAR 5. Transfer Entropy 6. Risk Index
Date of First Compliant Deposit:	15 May 2026
Last Modified:	15 May 2026 12:41
URI:	https://orca.cardiff.ac.uk/id/eprint/187016

Actions (repository staff only)

Edit Item