In this paper, we estimate the backtestable version of the Conditional Value-at-Risk (i.e. CoVaR≤) by fitting different multivariate parametric models that capture four stylized facts about multivariate financial time series of equity returns: heavy tails, negative skew, asymmetric dependence, and volatility clustering. While the volatility clustering effect is got by AR-GJRGARCH dynamics, the other styliized facts are captured through both non-Gaussian multivariate models and copula functions. The CoVaR≤ is computed for four Italian assets on the basis of the multivariate normal, the multivariate normal tempered stable model, the multivariate generalized hyperbolic model and the AIC-based best copula function among a set of Elliptical and Archimedean copulas.
CoVaR and backtesting: a comparison between a copula approach and parametric models
de luca giovanni
;rivieccio g;
2020-01-01
Abstract
In this paper, we estimate the backtestable version of the Conditional Value-at-Risk (i.e. CoVaR≤) by fitting different multivariate parametric models that capture four stylized facts about multivariate financial time series of equity returns: heavy tails, negative skew, asymmetric dependence, and volatility clustering. While the volatility clustering effect is got by AR-GJRGARCH dynamics, the other styliized facts are captured through both non-Gaussian multivariate models and copula functions. The CoVaR≤ is computed for four Italian assets on the basis of the multivariate normal, the multivariate normal tempered stable model, the multivariate generalized hyperbolic model and the AIC-based best copula function among a set of Elliptical and Archimedean copulas.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.