Beruflich Dokumente
Kultur Dokumente
ARCH Model
Garch Models
Forecasting Volatility using Garch Models
Results
Dataset, Software and References
Sumit Sourabh
Ravi Ranjan Singh
Sheetanshu Gupta
Sahil Singhal
Introduction
ARCH Model
ARCH Model
According to the simple Autoregressive Conditionally
Heteroskedastic ARCH(m) model the variance is given by
m
X
2 2
σn = γVL + αi un−i
i=1
m
X
where γ + αi = 1, VL is the long term variance rate , γ
i=1
and αi0 s are the respective weights assigned.
We can use Generalised Least Squares or maximum likelihood
estimation to estimate the ARCH models.
Defining ω = γVL the above equation can be rewritten as
Xm
σn 2 = ω + 2
αi un−i
i=1Series Analysis of GARCH Models for Volatality
Sumit Sourabh Ravi Ranjan Singh Sheetanshu Gupta Sahil Singhal
Time
Introduction
ARCH Model
Garch Models The Garch Model
Forecasting Volatility using Garch Models Estimation of Garch Models
Results
Dataset, Software and References
σn 2 = γVL + αun−1
2 2
+ βσn−1
Plot of Returns
We used the end of the day BSE index for the last couple of years
Plot of ACF
Plot of PACF
The end of day data for the stocks listed in National Stock
exchange is freely available at http:\\finance.yahoo.com.
For estimating the ARCH and GARCH model parameters we
use the MATLAB Garch toolbox which uses iterations to
maximize the maximum likelihood fucntion.
References
Thank You