Multiscale dependence analysis and portfolio risk modeling for precious metal markets

In this paper, we propose a new Bivariate EMD copula based approach to analyze and model the multiscale dependence structure in the precious metal markets. The proposed model constructs the Copula based dependence structure formulation in the Bivariate Empirical Mode Decomposition (BEMD) transformed multiscale domain. We further propose the BEMD Copula based Portfolio Value at Risk (PVaR) model to estimate the precious metal market risk measure. Empirical studies in the typical precious metal markets have been conducted. We found the evidence of multiscale structure of the time varying dependence structure among precious metal markets. We show that significantly improved portfolio risk forecasting performance could be achieved with the proposed model when the multiscale dependence structure is taken into account during the modeling process. •We propose BEMD model to analyze and model multi scale portfolio risk structure.•We incorporate the heterogeneous market risk structure information.•Normal and transient portfolio risk factors are separated and modeled individually.•We analyze the dependence structure between Gold and other precious metal markets.•The proposed model achieves the statistically significant improved performance.