Modeling Indonesian Crude Price Using Markov Switching Generalized Autoregressive Conditional Heteroscedasticity

  • Ria Retna Herliyasari, Nur Iriawan, Heri Kuswanto

Abstract

Indonesian Crude Price (ICP) is one of the important factors in determining the state
budget. If ICP prediction is far from its realization value, it will give a risk to state
revenue. The risks given by ICP predictions can be minimized by learning the
characteristics and patterns of ICP data. Like other financial time series data, the crude
oil price has a nonlinear pattern and fairly high volatility so that it has a nonconstant
variance (heteroscedasticity). Moreover, ICP data also have a structural change.
Previous studies showed that Generalized Autoregressive Conditional Heteroscedasticity
(GARCH) can handle the heteroscedasticity property in the data. However, GARCH can
not capture the structural change. Markov Switching is an alternative of time series data
modeling having structural change. In Markov Switching, structural change is considered
as a random event. In this study, the combination of GARCH and Markov Switching
called Markov Switching GARCH (MS-GARCH) was applied to model Indonesian Crude
Oil Price and compare it to the single regime GARCH model. Both of volatility models
are obtained under normal and Student’s t distribution. The best model is chosen based
on the smallest Akaike Information Criterion (AIC). According to AIC values, Markov
Switching GARCH gives better performance than single regime GARCH and Student’s t
distribution beat the normal distribution.

Published
2020-05-01
How to Cite
Ria Retna Herliyasari, Nur Iriawan, Heri Kuswanto. (2020). Modeling Indonesian Crude Price Using Markov Switching Generalized Autoregressive Conditional Heteroscedasticity. International Journal of Advanced Science and Technology, 29(7s), 3272-3279. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/17608