Archimedean copulae for risk measurement

Abstract

In this paper some Archimedean copula functions for bivariate financial returns are studied. The choice of this family is due to their ability to capture the tail dependence, which is an association measure we can detect in many bivariate financial time-series. A time-varying version of these copulae is also investigated. Finally, the Value-at-Risk is computed and its performance is compared across different copula specifications.

Keywords:

• copula
• time-varying parameters
• daily equity returns
• risk management
• value-at-risk

Notes

We thank a referee for pointing out that on the right-hand side of Equation (4) the logit of could replace . The same comment also applies to Equation (5).