Vine Copula Model and VaR Forecast for Multi-Asset Portfolio
Jiang Gao
Abstract
Jiang Gao
Abstract
The risk management of portfolio usually involved in more than two assets,under the condition that traditional bivariate Copula faces the problem of 'dimensional curse' and multivariate Copula functions have limitation on precise and flexibility when they are used to characterize the multivariate joint distribution.Vine-Copula is introduced to characterize multi-asset returns' joint distribution,based on different Pair-Copula decomposition,Vine-Copula is constructed.Then VaR of the portfolio is calculated by Monte Carlo method and the VaR forecast efficiency of Vine Copula model is backtested through Kupiec and Christoffersen backtesting and compared with traditional variance-covariance risk management method.Conclusion could be obtained that traditional variance-covariance method and Vine Copula based on normal Pair Copula decomposition method can not forecast VaR of multi-asset portfolio successfully.The Vine-Copula based on student-t pair Copula and Clayton Copula which show tail distribution feature can pass the backtesting.The conclusion can be used in guiding practice better.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The risk management of portfolio usually involved in more than two assets,under the condition that traditional bivariate Copula faces the problem of 'dimensional curse' and multivariate Copula functions have limitation on precise and flexibility when they are used to characterize the multivariate joint distribution.Vine-Copula is introduced to characterize multi-asset returns' joint distribution,based on different Pair-Copula decomposition,Vine-Copula is constructed.Then VaR of the portfolio is calculated by Monte Carlo method and the VaR forecast efficiency of Vine Copula model is backtested through Kupiec and Christoffersen backtesting and compared with traditional variance-covariance risk management method.Conclusion could be obtained that traditional variance-covariance method and Vine Copula based on normal Pair Copula decomposition method can not forecast VaR of multi-asset portfolio successfully.The Vine-Copula based on student-t pair Copula and Clayton Copula which show tail distribution feature can pass the backtesting.The conclusion can be used in guiding practice better.
Key concepts: Vine copula, Copula (linguistics), Joint probability distribution, Econometrics, Portfolio, Multivariate t-distribution, Multivariate statistics, Marginal distribution