Asymmetric Dependence Implications for Extreme Risk Management

作者：Tsafack Georges^*

来源：Journal of Derivatives, 2009, 17(1): 7-20.

DOI：10.3905/JOD.2009.17.1.007

The stronger dependence of lower returns compared to the dependence of upper returns in financial markets is a well-established stylized fact known as asymmetric dependence. This article analyzes the implications of this dependence structure on the measurement of extreme risk. Defining a new ordering concept related to the lower extreme dependence, this article shows that in the presence of asymmetry in the tail dependence, a portfolio model based on a multivariate symmetric model like DCC with Gaussian or Student-t innovations will lead to an underestimation of the portfolio value at risk (VaR) or expected shortfall (ES). The article develops an application to equity and bond portfolios and finds that at the 5% level, symmetric DCC models provide as good an estimation of VaR as the asymmetric dependence models, but a poor estimation of ES. For a more prudential level like 1% or 0.5%, the symmetric DCC models tend to underestimate VaR more than ES. However, when the Gumbel model, which displays an asymmetric dependence, is used, very accurate results at different levels are obtained.

• 出版日期2009