Page 1 of 1

GARCH-X model

Posted: Fri Mar 15, 2013 5:32 pm
by G_Cass
hello, i am obtaining a negative constant for my conditional variance
equation, please can anyone help? I am running a garch-x model on eviews, using the arch-garch object-equation option. my data is the uk 3-month t-bill. the garch-x
is very similar to the garch, except it include an additional regressor term in the variance equation, which is the last time periods interest rate.. I also have a dummy variable in my mean
equation.
my equation is defined as follows:
the mean equation, i type in: DELTA_UK=C(1)+C(2)*UK_2GAMMA_0P5+C(3)*DUMMY_CRASH
variance regressors (additional to what eviews suppiles) are:
UK_2GAMMA_0P5 DUMMY_FUTURES
i choose arch 1 and garch 1
i select the BHH method and derivative calculated on accuracy via the options
then i click ok
however the output gives a negative constant in the variance equation,
which implies that i can get a negative variance when the other
regressors are zero.

please can any one help resolve this, i am willing to post my eviews workfile too if required.

thank you very much, in advance.