Dynamic Model of Forecasting Stock Prices

Abstract

Sharia based investments currently become more popular in Indonesia as an alternative for those who have a long-term horizon and are seeking an Islamic way in investing their money. However, such long-term investment allows the existence of heteroscedasticity or heterogeneous variances in the time series data. To come up with this issue, one way to model the Autoregressive Conditional Heteroscedasticity (ARCH) effect is GARCH Model. The objective of this study is to obtain the best model estimating the parameters, to forecast the stock prices and to present its predicted volatility. The results show that the best model as fitted data is AR (1)-GARCH (1,1). The implication of this model is to predict the share price of Indofood CBP Sukses Makmur Tbk, Indonesia, for the next 2 months (60 days) and it shows a very reasonable result as the percentage of error is less than the mean.