Skip to main content

Modeling and Testing the Evolution of Systemic Risk and Herding Behavior in Financial Markets

Vioto, Davide (2019) Modeling and Testing the Evolution of Systemic Risk and Herding Behavior in Financial Markets. Doctor of Philosophy (PhD) thesis, University of Kent,. (Access to this publication is currently restricted. You may be able to access a copy if URLs are provided) (KAR id:80377)

PDF Other
Language: English

Restricted to Repository staff only until 31 December 2022.
Contact us about this Publication
[thumbnail of 154PhD_Dissertation_DV.pdf]

Abstract

This research investigates: i) the evolution and the information content of market-based systemic risk measures (SRM); ii) the main drivers of herding behavior in equity markets; and, iii) the existing relationship between the market-based risk measures used to estimate systemic risk and herding behavior.

Chapter 2 presents a detailed and extensive literature review on which this thesis is based. It explores the multi-disciplinary approaches to analysing systemic risk and herding behavior, the data requirements and the main measures defined in the existing financial literature.

In Chapter 3, we study the effect of estimation uncertainty of market-based SRMs on selecting and regulating global systemically important banks (G-SIB). Using the three leading SRMs, we test how closely they agree with the list of G-SIBs from the Financial Stability Board (FSB) and how closely the SRMs match the categorization of G-SIBs into the five systemic risk buckets used by the FSB to assign capital surcharges to G-SIBs. Second, using cluster analyses we provide an alternative procedure to identify G-SIBs based on SRMs. This procedure incorporates the SRM confidence intervals of banks and is used to assess the degree of prudence versus conservatism that the FSB applies in compiling their G-SIBs list. Third, our approach integrates the SRM confidence interval in assigning a G-SIB to a systemic risk bucket and in determining the capital surcharge of each bucket. In general, we find that the three SRMs collectively are efficient in discriminating between systemic and non-systemic banks. The systemic risk buckets defined by the FSB are different from those constructed in a full pairwise comparison approach based on the market measures. In addition, we identify banks that were not marked as systemically important by the scoring method of the FSB but that are systemically important based on market-based SRMs. Finally, as the ranking with SRMs is subject to risk estimation uncertainty, we show how the ranking process can be improved by employing confidence intervals. Our methodology is able to identify as systemically important the banks designated as G-SIBs through supervisory judgment by the financial authority. The results also show that a G-SIBs designation based only on SRM point estimate would assign higher additional capital buffers compared to our new method.

Chapter 4 aims to contribute to the debate on systemic risk by assessing the level of systemic risk of China's financial system over the period from January 2010 to December 2016, a period spanning the deflation of China's property bubble, the banking liquidity crisis, and the stock market crash. We focus on China's financial system because it became a source of risk during the banking liquidity crisis of 2013, and concerns regarding the level of systemic risk of the financial system increased following the popping of the stock market bubble in the summer of 2015. Dividing the financial system into three sectors, namely: banks, insurance and brokerage industries, and real estate, and applying the ΔCoVaR as the measure for systemic risk; our findings show that the systemic risk level of China's financial system decreased following the deflation of the property bubble in 2012, and successively increased during the banking liquidity crisis in 2013, reaching a major peak during the market crash in 2015. We further show, through the Wilcoxon signed rank test, that the systemic risk level of the financial system and sectors significantly increased after the main systemic events. In order to provide a formal systemic risk ranking of the financial sectors, we apply the bootstrap Kolmogorov-Smirnov test, finding that each financial sector significantly contributes to systemic risk, with the banking sector contributing the most, followed by real estate and subsequently insurance and brokerage industries.

In Chapter 5, we provide new evidence of herding in global markets. Using OLS and quantile regressions and applying daily data for 33 countries from January 2000 to end-January 2019, we find evidence of herding in few Asia Pacific, Latin American and European markets. When, however, we condition on the Eurozone crisis and the China's market crash of 2015-16, we find significant evidence of herding for most countries. We also document important herding behavior evidence related to Brexit. This Chapter pioneers research on the relationship between herding and systemic risk. By conditioning the investigation of herding behavior on different systemic risk levels of the market, the results strongly suggest that herding is more pronounced in case of high systemic risk. Granger causality tests and Johansen's vector error-correction model provide solid evidence of the existence of a strong relationship between herding and systemic risk, suggesting that herding behavior may be an ex-ante aspect of systemic risk.

Lastly, Chapter 6 tests for herding towards the market consensus for the corporates in the United States and the Eurozone equity markets, considering their main financial industries. We find that herding is more likely to be present in the high quantiles, entailing herding effects under turbulent market conditions. This effect appears more pronounced when we condition on the financial crises periods. Our results also support the herding presence in case of asymmetric conditions of volatility, credit deterioration, funding illiquidity and economic policy uncertainty. Furthermore, we provide evidence that the cross-sectional dispersion of returns of the domestic equity market can be partly explained by the corresponding dispersions of the financial sector and its industries, with the latter having influence on the herding of the domestic equity market. In our analysis we cover the last two main global financial crises, revealing new evidence of "spurious" and "intentional" herding corporate activity.

Chapter 7 provides final discussions, concluding the thesis and points towards new directions in which this research might go in order to fill gaps still open in literature, such as: i) cross-country contagion in financial markets due to systemic risk; ii) the use of market-based SRMs as risk measures in risk-budgeting/parity portfolios; iii) the investigation of capital requirements as instrument to assess systemic risk; iv) the examination of the interrelationship between systemic risk and climate-change risks in the insurance sector; and, v) the herding investigation in the option market.

Item Type: Thesis (Doctor of Philosophy (PhD))
Thesis advisor: Tunaru, Radu
Thesis advisor: Morelli, David
Uncontrolled keywords: Systemic risk, Herding behavior, Confidence intervals, Financial stability, Capital requirements, Quantile regressions, Dominance test
Subjects: H Social Sciences > HF Commerce > HF5351 Business
Divisions: Divisions > Kent Business School - Division > Kent Business School (do not use)
SWORD Depositor: System Moodle
Depositing User: System Moodle
Date Deposited: 06 Mar 2020 13:10 UTC
Last Modified: 16 Feb 2021 14:11 UTC
Resource URI: https://kar.kent.ac.uk/id/eprint/80377 (The current URI for this page, for reference purposes)
  • Depositors only (login required):