The University of Southampton
University of Southampton Institutional Repository

Portfolio based VaR model: a combination of extreme value theory (EVT) and dynamic conditional correlation (DCC) model

Portfolio based VaR model: a combination of extreme value theory (EVT) and dynamic conditional correlation (DCC) model
Portfolio based VaR model: a combination of extreme value theory (EVT) and dynamic conditional correlation (DCC) model
This thesis fills a gap in the risk management literature and expands the understanding of the portfolio value at risk (VaR) by providing a theoretical market risk measurement of a portfolio (called “GEV-DCC model”), which combines the tail dynamic conditional correlation (tail-DCC) and extreme value theory. According to the spirit of VaR, the tail distribution is more important than the entire distribution, as well as the correlation in the tail area between various assets. The main advantage of this approach is the increase of accuracy in the parameter estimation of the tail distribution and more consistent correlation measurement for VaR. The results from this method are compared with four other conventional VaR approaches; GARCH model, RiskMetrics, stochastic volatility, and historical simulation. Furthermore, three quality measures are applied to evaluate the suitability, conservativeness, and magnitude of loss of the forecasted VaR, which offer more information from the forecasted VaR pattern.

Applying 16 major equity index returns from developed and emerging markets, this study finds that the GEV-DCC model offers a more accurate coverage across the blocks in the three hypothetical portfolios (the developed equity markets, Asian and Latin American equity markets) compared with the four competing models. The uncovered rates of the GEV-DCC model with the 5-day block approach are generally close to the given probability (?) set in the VaR calculation.

These consistent results can also be found in the robustness test with the shorter forecasting period. In the quality checks, the GEV-DCC presents a relatively stable pattern in the daily and 10-day VaR results. In addition, the GEV-DCC model also provides satisfactory results in the conservativeness and potential loss tests although no direct evidence indicates that it delivers the best result in these two checks. We also find significant differences between the original DCC and the tail-DCC. This evidence shows that the correlations between equity markets in the left tail are significantly higher than the ones in the right tail, and there are significant changes (generally rising) in the tail-DCC patterns around the period of financial crisis in the third quarter of 2008.

The results from this study could potentially provide a critical reference for investors in measuring or managing the market risk.
Wang, Jo-Yu
edffb9e5-0b97-474d-acbb-44cd81da6cc6
Wang, Jo-Yu
edffb9e5-0b97-474d-acbb-44cd81da6cc6
Choudhry, Taufiq
6fc3ceb8-8103-4017-b3b5-2d38efa57728

(2013) Portfolio based VaR model: a combination of extreme value theory (EVT) and dynamic conditional correlation (DCC) model. University of Southampton, School of Management, Doctoral Thesis, 332pp.

Record type: Thesis (Doctoral)

Abstract

This thesis fills a gap in the risk management literature and expands the understanding of the portfolio value at risk (VaR) by providing a theoretical market risk measurement of a portfolio (called “GEV-DCC model”), which combines the tail dynamic conditional correlation (tail-DCC) and extreme value theory. According to the spirit of VaR, the tail distribution is more important than the entire distribution, as well as the correlation in the tail area between various assets. The main advantage of this approach is the increase of accuracy in the parameter estimation of the tail distribution and more consistent correlation measurement for VaR. The results from this method are compared with four other conventional VaR approaches; GARCH model, RiskMetrics, stochastic volatility, and historical simulation. Furthermore, three quality measures are applied to evaluate the suitability, conservativeness, and magnitude of loss of the forecasted VaR, which offer more information from the forecasted VaR pattern.

Applying 16 major equity index returns from developed and emerging markets, this study finds that the GEV-DCC model offers a more accurate coverage across the blocks in the three hypothetical portfolios (the developed equity markets, Asian and Latin American equity markets) compared with the four competing models. The uncovered rates of the GEV-DCC model with the 5-day block approach are generally close to the given probability (?) set in the VaR calculation.

These consistent results can also be found in the robustness test with the shorter forecasting period. In the quality checks, the GEV-DCC presents a relatively stable pattern in the daily and 10-day VaR results. In addition, the GEV-DCC model also provides satisfactory results in the conservativeness and potential loss tests although no direct evidence indicates that it delivers the best result in these two checks. We also find significant differences between the original DCC and the tail-DCC. This evidence shows that the correlations between equity markets in the left tail are significantly higher than the ones in the right tail, and there are significant changes (generally rising) in the tail-DCC patterns around the period of financial crisis in the third quarter of 2008.

The results from this study could potentially provide a critical reference for investors in measuring or managing the market risk.

PDF
Final PhD thesis - Jo-Yu Wang.pdf - Other
Download (8MB)

More information

Published date: January 2013
Organisations: University of Southampton, Southampton Business School

Identifiers

Local EPrints ID: 348328
URI: http://eprints.soton.ac.uk/id/eprint/348328
PURE UUID: cc4ea1a8-d8da-4f76-ad14-1bad38427b2b
ORCID for Taufiq Choudhry: ORCID iD orcid.org/0000-0002-0463-0662

Catalogue record

Date deposited: 28 Feb 2013 15:07
Last modified: 24 Jul 2019 00:36

Export record

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×