Generalized Auto-Regressive Conditional Heteroskeasticity (GARCH) is a model used to predict the volatility of returns. Volatility is a statistical measure of the movement of returns for securities (financial instruments that can only be traded through markets or securities companies) or certain market indices. Then the GARCH model was further developed into an asymmetric form, namely conditional volatility and returns have a relationship, namely the GJR model which is an abbreviation of the name (Glosten- Jagannathan-Runkle). This research focuses on the GJR-X by adding high-frequency exogenous variables in volatility process and on the GARCH-CJ which is a decomposition of the exogenous variable X, namely the continuous component C (Continuous) and the jump J (Jump). TOPIX data (Tokyo Stock Price Index) is the real data used in this study. To estimate the model parameters, the ARWM (Adaptive Random Walk Metropolis) method will be used with the MCMC (Markov Chain Monte Carlo) algorithm. First, it was found that the ARWM method is good at estimating parameters. Second, the AIC value of GJR-CJ was smaller than that of GJR-X, which means that GJR-CJ had better data fitting.
Copyrights © 2023