Page 1 of 1

state space model complicacy.

Posted: Mon Dec 14, 2009 2:22 pm
by Amazing
This is a state space model which has been structured to estimate the potential GDP, Natural rate of interest and Trend growith rate and other set of exogenous factors (all unobservables) under the light of Laubach and Williams model 2003
@signal 100*(y-sv1)=c(1)*100*(y(-1)-sv1)+c(2)*(r(-1)-sv2)+[var=exp(c(3))]
@signal p=c(4)*p(-1)+c(5)*100*(y(-1)-sv1)+[var=exp(c(6))]
@state sv2=c(7)*sv3+sv4
@state sv4=c(8)*sv4(-1)+[var=exp(c(9))]
@state sv1=sv1(-1)+sv3(-1)+[var=exp(c(10))]
@state sv3=sv3(-1)+[var=exp(c(11))]
Where y is the output
r is the real interest rate

In this model (Y-SV1) represents the output gap which is a function of its own lag and the lagged difference between the real interest rate and natural rate of Interest. The first equation being aggregate demand equation and the Second being aggregate supply equation.
In this model SV1 is the potential GDP, Sv2 is the Natural rate of interest, and SV3 is the trend growth rate and SV4 is the other set of exogenous factors. The problem is Eviews is not helping me out in estimating the model by giving two common errors
SV1 is not defined
SV4 is not allowed
Can somebode come up with a solution to this? can someone guide me through this that what's wrong in this specification?

Re: state space model complicacy.

Posted: Mon Dec 14, 2009 3:59 pm
by EViews Glenn
There are a number of problems with the specification. While there may be more, two stand out at a glance.

1) You can't have SV1 on the left-hand side of a signal equation.

2. The SV2 equation has to be written in terms of lagged states

Re: state space model complicacy.

Posted: Tue Dec 15, 2009 11:44 am
by Amazing
Thanks for entertaining my problem. But let me present the actual model for which i represented that state space model. The attached file is placed inline

Re: state space model complicacy.

Posted: Tue Dec 15, 2009 11:59 am
by Amazing
one more thing i forgot to mention that zt being exogenous unobserved variable so how can i write in lagged form. anyways i have sent you the original model, i would be very grateful if you show me how this whole model be represented in state space form if my representation is inaccurate.

Re: state space model complicacy.

Posted: Tue Dec 15, 2009 2:53 pm
by EViews Glenn
Translating user specifications into state space form is beyond our normal technical support. Perhaps someone else on the forum can help.

I will note that you have an rt* equation which is a simple linear combination of contemporaneous states. Note that you can easily replace this in your specification with the constituent states.