Volatility model estimations of palm oil price returns via long-memory,asymmetric and Heavy-Tailed GARCH parameterization

This study attempts to model the volatility of palm oil price returns via a number of generalized autoregressive conditional heteroskedasticity class of models that capture the long-range memory,asymmetry,and heavy-tailedness phenomena. these models have been estimated in the presence of four alternative conditional distributions: gaussian,student t,generalized error distribution,and skewed student t. the empirical results indicate that complex model specifications and distribution assumptions do not seem to outperform the simpler ones in terms of standard model selection criteria and numerical convergence. with regard to the conditional distributions,a symmetric fattailed distribution has been found to be preferred to gaussian and asymmetric distribution in many cases.