PENGGUNAAN METODE VALUE at RISK UNTUK MENENTUKAN TINGKAT RESIKO INVESTASI PADA SAHAM PT GUDANG GARAM Tbk MELALUI PENDEKATAN MODEL INTEGRATED GENERALIZED AUTOREGRESSIVE CONDITIONAL HETEROSCEDASTICITY (IGARCH)

Abstract: Heteroscedasticity in most economic and financial time series data can be overcome  by  ARCH  /  GARCH models.  However,  the  application  of  ARCH  / GARCH models has several weaknesses, one of  them is the inability to  see the change or transition of behavior between low volatility and high volatility. Hence, ARCH  /  GARCH  models  are  modified  by  using  IGARCH  models. In  the IGARCH  model,  stationarity  is  satisfied  if sum  squares of  residual coefficients and  conditional  variance  is  1. IGARCH  model  is better than ARCH  /  GARCH models.  The  purpose  of  this  paper  is  to  model  the  stock  price  index that  is heteroscedastic into IGARCH model and to determine the Value at Risk (VaR) of stock  price  index  for a  period  ahead.  The  data used  in  this  paper is  stock  price return data of PT Gudang Garam Tbk.
Modeling  is settled by  forming  ARIMA  model  as  a mean model, and followed by modeling ARCH / GARCH, and then modeling conditional variance IGARCH  model  where the  sum  of both  parameters  IGARCH  number  of coefficients equal to one. From the stock price data of PT Gudang Garam Tbk, we obtained  IGARCH (1, 1) as the best model because it satisfied all assumptions, those are, parameters which had been significant and squared residuals which had been white noise. The general form of the IGARCH (1, 1) model is = 3.87 × 10 + 0.104046 + 0.895954ℎ . Value  at  Risk (VaR) is  obtained  from the formed model for the next 30 days with certain allocated funds. Therefore, if the allocated funds amounts IDR 1 billion with error rate of 95%, the level of risk faced by investors who will invest in companies of PT Gudang Garam Tbk will be IDR 210.420.382,00.
Keywords: ARCH, GARCH, IGARCH, Value at Risk, Volatility, Return, White Noise, Heteroscedasticity
Penulis: Nasrudin MB, Sediono, Eko Tjahjono
Kode Jurnal: jpmatematikadd130073

Artikel Terkait :