BAYESIAN ESTIMATION OF THE GJR-GARCH (p, q) MODEL WITH STUDENT-T PRIOR DISTRIBUTION

The presence of volatility in many financial time series data is one of the problems that cause the variance to be non-constant. The GJR-GARCH (p, q) is a model that takes into account time-varying volatility, allowing positive and negative shocks to have distinct effects. This study provides the estimates of the GJR-GARCH (p, q) model using the Bayesian approach. Student-t distribution is used as prior error distribution. It derives the posterior distribution of the GJR-GARCH (p, q) model with student-t distribution, specifically the parameters α and β, latent variable ω, and degrees of freedom v.