Backtesting VaR in consideration of the higher moments of the distribution for minimum-variance hedging portfolios

The higher moments of a distribution often lead to estimated value-at-risk (VaR) biases. This study's objective is to examine the backtesting of VaR models that consider the higher moments of the distribution for minimum-variance hedging portfolios (MVHPs) of the stock indices and futures in the Greater China Region for both short and long hedgers. The results reveal that the best backtesting VaR for the MVHP considered both the higher moments of the MVHP distribution and the asymmetry in volatility, cross-market asymmetry in volatility, and level effects in the covariance matrix of assets in the MVHP. These empirical results provide references for investors in risk management.