COMPARISON STUDY BETWEEN GARCH AND STANDARD DEVIATION

Abstract

In this research, the implementation of GARCH Model to Sharpe formula to replace standard deviation . In this research, the analysis data of Adaro Energy Tbk, Adhi Karya Tbk and Astra Agro Lestari Tbk from the Indonesia LQ45 are studied by using analysis time series AR(p)-GARCH(p,q) modeling. From the analysis, it was found that the data Adaro Energy Tbk, Adhi Karya Tbk and Astra Agro Lestari Tbk are nonstationary. To make the data stationary, the differencing process with lag=2 (d=2) is used and the time series data then attain stationary. From the test of ARCH effects by using Q test and Lagrange Multiplier it concludes that all the data Adaro Energy Tbk, Adhi Karya Tbk and Astra Agro Lestari Tbk have ARCH effects. Based on this situation, then the AR(p)-GARCH(p,q) model are used to modeling the data.The best model for all data Adaro Energy Tbk, Adhi Karya Tbk and Astra Agro Lestari Tbk are the AR(1)-GARCH(1,1) models. The result that the models AR(1)-GARCH(1,1) for data Adaro Energy Tbk, Adhi Karya Tbk and Astra Agro Lestari Tbk are very fit with the data and based on the criteria MAPE (The mean absolute percentage error). It shows that the forecasting are very reliable. From the analysis we can get that GARCH is better than standard deviation witf differences 2,9% for PT Adhi Karya Tbk, -0,5% for PT Astra Agro Lestari Tbk, and -3,7% for PT Astra Agro Lestari.