EViews.com

I've a couple of questions on kalman filter, which I am using to compute for natural rate of return.

1. I wanted to run a kalman filter with the following specs:

@signal y = SV1 + error var
@signal z = bunch of exogenous vars + c( )*SV1 + error var
@state SV1 = SV3 + SV1(-1) + error var
@state SV2 = SV2(-1) + error var
@state SV3 = SV2(-1)

my concern is the SV3 in state equation 1. can eviews handle this? basically, i have a 2nd state var in the first state equation.


2. I wanted to run a kalman filter with the following specs:

@signal y = c( )*SV1 + SV4 - SV5 + error var
@signal m = SV5 + bunch of lags of m + error var
@signal SV1 = SV1(-1) + error var
@signal SV2 = SV1(-1)
@signal SV3 = SV2(-1)
@signal SV4 = c( )* SV2 + c( )*SV3 + c( )*SV5 + c( )* SV6 + bunch of exogenous variables + c( )*m(-1) + c( )*m(-2) +c( )*m(-3) + error var
@signal SV5 = SV4(-1)

my concern is the bunch of lags of m in state equation for SV4. can eviews handle this? if not, what is the work-around?

Thanks and appreciate your advice/tips.

Hi all

I would really, really, really be grateful for help on the query below...

I want to use eviews for my research, but I am getting feedback from colleagues that the 2 cases below would require me to use Matlab or R. But as I've said, I prefer eviews. So if anyone knows any work-around, can you please share?

Thank you very much.

As it says in the Help system

Each state equation must be linear in the one-period lag of the states. Nonlinearities in the states, or the presence of contemporaneous, lead, or multi-period lag states will generate an error message. We emphasize the point that the one-period lag restriction on states is not restrictive since higher order lags may be written as new state variables

Thanks for the reply, startz. I guess I was hoping there was a work-around similar to the work around for higher order lags. Let me know if such work-arounds exist. Thanks!