GARCH Instruction problems

[sulong]

eq.jpg (39.53 KiB) Viewed 5532 times

Hi Tom ,
I am trying to use RATS to estimate the coefficients for the above GJR Garch specification, used by Bohl & Brzeszczynski, 2006 in the attached article.

But I am facing some difficulties with the instructions,
The following is what i have so far, but how about

1.jpg (2.93 KiB) Viewed 5532 times

and

2.jpg (3.55 KiB) Viewed 5532 times

how can i instruct rats to include the above coefficients in the following code?

additionally, how about the Xregressors in the variance equation.


GARCH(P=1,Q=1,XREGRESSORS,REGRESSORS,Pmethod=sim,piters=20,resids=u,hseries=h,distrib=1,asymmetric,i=drift,method=1) / %S('r'+%l(i))
# Constant DoW1 DoW2 DoW3 Dow4 DoW5 DDoW1 DDoW2 DDoW3 DDow4 DDoW5 R(subscript t-1) (D(subscript)R(subscriptt-1)) R(subscript f)
#

[TomDoan]

For many reasons, you can't use GARCH to estimate that---there are just too many effects that have to be generated as part of the recursion. You'll have to use a MAXIMIZE set up. You should look at the replication file for Baillie and Bollerslev, "The Message in Daily Exchange Rates: A Conditional Variance Tale", JBES 1989, vol 7, pp 297-305 which uses day of the week effects.

[sulong]

Ok Tom, will do.

thanks