Risk modelling using Vine Copulas: Modelling an energy company portfolio
MetadataShow full item record
In this paper, a method for calculating Value-at-Risk using GARCH and Vine Copulamodelling with various marginals is implemented and tested on a set of eight electricity futures. The forecasts from this model are then compared to similar forecasts using a DCC-GARCH-model, RiskMetrics and historical simulation. These are all compared using the Kupiec and Christophersen tests. The comparison showed that at the 1%- and 99%- quantiles the Vine Copula method performs best, and the GARCH-based models generally outperformed the others. The Vine Copula performed worse than the benchmark models at the 5%- and 95%-quantiles. DCC-GARCH was able to predict all the quantiles fairly well in most of the portfolios.