Page 195 - Special Topic Session (STS) - Volume 1

STS425 Arifah B. et al.
This industry is important and vital in Malaysia and consequently any change
in crude oil prices will affect the country’s earning indirectly and will form a
risk for all the workers in this industry. Therefore, forecasting the CPO price is
important to make ease decision in the immense economic instability event
and to plan for various investment and business activities for optimal resource
allocation. The high degree of accuracy of the predictions of CPO price is of
paramount importance since any decisions will be taken based on this
predictions will affect the performance of the true market of this commodity.
This task is not an easy one where it includes long memory stochastic volatility
(LMSV). For improving the accuracy of CPO price forecasting, it is necessary to
monitor the CPO price and to record its growth data for long time period.
There are many stochastic volatility (SV) models have been developed in
literature to forecast the CPO price. Arshad and Zainalabidin (1994) examined
the ability of CPO future market to predict the forward prices efficiently. They
found that the futures price method outperforms the other techniques such
as moving average, Box Jenkins, exponential smoothing and econometric in
forecasting the forward CPO price. One of the most popular model is the Box
and Jenkins model that produces the results for linear time series data (Karia
et al., 2013b). Arshad and Ghaffar (1986) forecasted the CPO price by using an
univariate autoregressive-integrated moving average (ARIMA) model which is
developed by Box Jenkins approach . Ahmad et al. (2014) applied ARIMA
model to find the suitable time series that can simulate monthly CPO price in
Malaysia. However, their residuals results were not normal and orthogonal.
Moreover, ARIMA models can give inaccurate estimations with large sample
sizes. Khin et al. (2013) compared between three statistical models (Vector
Error Correction Method (VECM), Multivariate Autoregressive Moving Average
(MARMA) and ARIMA model) that have been used to forecast the spot palm
oil price in Malaysia. Their results demonstrated that MARMA model is the
superior in comparison with VECM and ARIMA models. However, the results
of the Root Mean Square Percentage Error (RMSPE) demonstrated that the
model give high percentage of errors. Omar and Majid (2004) used the
historical variances returns of spot and futures price to investigate the
relationship between the spot and futures prices of CPO contracts that are
traded in the Malaysian Derivatives Exchange.
All the previous studies predict the price of CPO for short memory
volatility. The discovery of long memory behavior in the volatility of some
financial data was started from the early 1990s. Ding et al. (1993) investigated
that there is strong correlation between absolute returns of the daily standard
and poor 500 index prices and they were among of the first people who
discover this relation. Due to the instability and volatility in CPO price, the
stochastics models such as autoregressive conditional heteroskedastic
(ARCH), generalized ARCH (GARCH), exponential GARCH (EGARCH) or

184 | I S I W S C 2 0 1 9