When aggregating financial risk on a portfolio level, the specification of the dependence structure between the risk factors plays an important role. Promising parametric models are often based on a so-called copula approach. Case studies of market crashes suggest the application of concepts allowing for extremal dependence. We present a transformed copula as a new model that both fits the data and allows for exact prediction in the tails. It turns out that the new model improves benchmark models like the t- or Clayton copula with respect to risk measures like VaR or Expected Shortfall. By performing different goodness-of-fit tests, the quality of the estimation is examined. Copyright 2005 Royal Economic Society

No abstract

SUMMARYIn this paper, we are concerned with bivariate di erentiable models for joint extremes for dependent data sets. This question is often raised in hydrology and economics when the risk driven by two (or more) factors has to be quantiÿed. Here we give a full characterization of polynomial models by means of their dependence function and dependence measure.

No abstract

No abstract