Tujuan penelitian ini adalah untuk mengukur dan mengetahui faktor-faktor yang mempengaruhi volatilitas pergerakan saham individual pada sektor infrastruktur, utilitas, dan transportasi sejak Januari 1998 sampai dengan Desember 2005. Model yang digunakan untuk mengetahui volatilitas tergantung dari hasil uji terhadap varian residual yang digunakan. Jika varian residual tersebut bersifat tidak konstan maka menggunakan model Autoregressive Heteroskedastic (ARCH) dan Generalized Autoregressive Heteroskedastic (GARCH). Namun bila varian dari residual konstan akan menggunakan model Ordinary Least Square, dalam penelitian ini model yang sesuai adalah AutoRegressive Moving Average (ARMA).Berdasarkan hasil pengolahan data selama periode penelitian ditemukan bahwa terdapat enam emiten yang volatilitasnya bersifat hornoscedastic, yaitu PT Berlian Laju Tanker (BLTA), PT Bukaka Utama (BUKK), PT Citra Marga Nusapala (CMNP), PT Petrosea (PTRO), PT Zebra (ZBRA) dan PT Telekomunikasi Indonesia (TLKM). Untuk kondisi demikian maka digunakan model AutoRegressive Moving Average (ARMA).Sementara terdapat enam emiten yang bersifat heteroseedastic yaitu PT Centris Multi Persada (CMPP), PT Humpus Intermoda (HITS), PT Indosat (ISAT), PT Rigs Tender (RIGS), PT Mitra Rajasa (MIRA) dan PT Steady Safe (SAFE), sehingga digunakan model ARCH dan GARCH. Selain itu, terdapat emiten yang memiliki ARCH/GARCH yang volatilitasnya bersifat persisten yang ditandai dengan nilai a + I yaitu RIGS (1998, 1999, 2000, 2001, 2002, 2004, 2005) HITS (1999), CMPP (2000). Adanya persistensi dalam volatilitas mengurangi kestabilan model ARCH/GARCH. Hal ini disebabkan sepanjang periode penelitian data return saham tersebut bersifat stagnan atau pergerakan return yang sangat tinggi dan tajam. The objectives of this research are to recognize and measure factors affecting the volatility of individual shares movement in infrastructure, utility and transportation sector during January 1998 - December 2005. Model used to identify volatility depend on the result of the test of the varian of residual on the used data. Should the varian of the residual act inconstantly then we use Autoregressive Heteroskedastic (ARCH) and Generalized Autoregressive Heteroskedastic (GARCH) model. But if the varian of the residual act constantly then we use Ordinary Least Square model, which in this research the model that fit the data is AutoRegressive Moving Average (ARMA) model.Base on the result of the data processing on the research period, it is found that there are six emitens whose volatility is homoscedastic, which are PT Berlian Laju Tanker (BLTA), PT Bukaka Utama (BUKK), PT Citra Marga Nusapala (CMNP), PT Pelrosca (PTRO), PT Zebra (ZBRA) dan PT Telekomunikasi Indonesia (TLKM). Therefore model used for these emitens is Auto Regressive Moving Average (ARMA) model.While there are six emitens whose volatility is heteroscedastic which are PT Centris Multi Persada (CMPP), PT Humpus lntermoda (HITS), PT Indosat (ISAT), PT Rigs Tender (RIGS), PT Mitra Rajasa (MIRA) dan PT Steady Safe (SAFE). Therefore model used for these emitens are either Autoregressive Heteroskedastic (ARCH) or Generalized Autoregressive Heteroskedastic (GARCH) model.It is also found within the research that there are six emitens whose models has a value of a + b ≥ 1, which depicts a persistency. They are RIGS (1998, 1999, 2000, 2001, 2002, 2004, 2005) HITS (1999), CMPP (2000). The presence of a persistency of a model may decrease the stability of the ARCHIGARCH model. This problem may occur due to the movement of the data which are very fluctuate or too stagnant. |