Page 1 of 1

GARCH-M (1,1) state space

Posted: Mon May 02, 2011 8:11 am
by Vasily
Hello

I need to estimate the following state space GARCH-M (1,1) model:

signal: r(t) = c(1 t) + c(2 t)*r(t-1) + h(t) + eps(t), eps~N(0,h(t))
state: h(t) = c(3) + c(4)*h(t-1) + c(5)*(eps(t-1))^2
state: c(i t) = c(i t-1) + v(i t), v(i t)~N(0, (sigma(i))^2), i=1,2

But there is an error in EViews that it can't estimate a state space model with nonlinear errors in state equation (which is for "h").
How can I make this estimation in EViews?

Best regards!

Re: GARCH-M (1,1) state space

Posted: Mon May 02, 2011 11:55 pm
by trubador

Re: GARCH-M (1,1) state space

Posted: Wed May 04, 2011 11:44 pm
by Vasily
@trubador

Thanks a lot!
But can you please give me some advices how could I linearize this model? I'm not so good in econometrics, but I really need to estimate this model. Here is my e-mail if this will be needed:

akimov.vn@mail.ru

Best regards!