Measuring portfolio risk of non-energy commodity using time-varying vine copula

In literature, many empirical studies have used the vine copula to measure and analyse the risk of stock indices, energetic products or crypto currencies portfolio (e.g., Zhang et al., 2014; Shahzad et al., 2018; Boako et al., 2019). These studies used mainly the VaR and ES with different versions of vine copula. However, few works have focused on a portfolio composed of non-energy commodities. Our aim in this work is to model the dependence between non-energy commodities returns by modelling standardised residuals obtained from univariate GARCH model by different versions of time varying vine copula, to quantify risk of N-dimensional non-energy commodity portfolio by VaR and ES and to compare the predictive performance of this method with traditional and competitive traditional univariate VaR methods. Empirical results suggest that risk quantifies generated by AR-GARCH vine copula methods with student-t distribution are sufficiently accurate at both low and high confidence levels. Given these results, we recommend the application of vine copula method to understanding the non-energy commodity behaviours which are very important to investors, producers, consumers, and policy makers.