Modelling crude oil-petroleum products’ price nexus using dynamic conditional correlation GARCH models

Modelling volatility in returns has continued to gain popularity with the evolution of the GARCH-type models under different frameworks. This study therefore examined the different variants of the multivariate GARCH model with focus on those that incorporated asymmetry and constant or dynamic conditional correlations. These variants were used in modelling the crude oil-petroleum products’ (gasoline, heating oil, kerosene, propane and diesel) price nexuses. Comparatively, the DCC-VAR-AMGARCH model fitted the return series more appropriately in four out of the five investigated nexuses, while the DCC-AMGARCH variant fitted the return series in just one nexus. With the exception of propane own market spillover, the overall volatility persistence of spillovers from own market and other markets for the nexuses of crude oil and the other four petroleum products (gasoline, heating oil, diesel and kerosene) were mean reverting. The study also adopted two hedging strategies, for each of the five crude oil-petroleum product nexuses, to ascertain plausible portfolio investment options. The empirical evidence on the different portfolio investments that were herein provided are especially useful for stakeholders/investors desiring to channel their resources into less risky investment portfolios.