Pemodelan Volatilitas Harga Saham GGRM dengan Pendekatan Generalized Autoregressive Conditional Heteroskedasticity (GARCH)

Abstract

This research aims to obtain the best model for the stock data of PT Gudang Garam Tbk. (GGRM) based on AIC and BIC values. The study utilizes secondary data from the Investing website, using daily stock data of GGRM for the past year. The initial steps involve testing data stationarity, identifying the ARIMA model, and checking for heteroskedasticity in the best ARIMA model. Subsequently, the ARCHGARCH model is identified. The study found that after modeling ARIMA for the stationary GGRM stock data, heteroskedasticity was present with a Ku=6, indicating a fat tail. Due to the presence of heteroskedasticity and fat tail, a GARCH model approach with a distribution assumption capable of handling the fat tail was applied. Considering the smallest AIC and BIC values, the best model identified is the T GARCH − Student T(1,1) with the equation o2t = 1.6986 + 0.1821 e2t-1 + 0.9020 o2t-1 − 0.1821e2t-1 It−1<0.