Tail Dependence of Major U.S. Stocks
RISK MANAGEMENT AND CORPORATE GOVERNANCE, Abol Jalilvand, Tassos Malliaris, ed., Routledge, 2011
Posted: 26 Aug 2011 Last revised: 29 Aug 2011
Date Written: August 25, 2010
Abstract
We review estimation methods of the tail dependence coefficient (TDC), simulating their finite-sample performance. With our chosen semi-parametric and non-parametric estimators, we estimate TDCs of major U.S. stocks. We have three aims. The first is to establish the “stylized facts” about tail dependence among major U.S. stocks. The second is to compare the “stylized facts” with TDCs implied by a multivariate Student’s t copula model so as to assess its ability of capturing tail dependence patterns of a large set of assets. The third is to explore the accuracy of estimates of VaR and Expected Shortfall when a multivariate Student’s t copula is used to capture tail dependence.
Keywords: copula, tail dependence, non-parametric estimation, risk measures, extreme value theory
JEL Classification: C14, C32.
Suggested Citation: Suggested Citation