Hello,
Is it possible solve system of eqautions with VAR?
I have two equations like;
Y = c(0) + c(1)X+ c(2)Z +c(3)M +c(4)N
X= c(5)+ c(6)Y+ c(7)P +c(8)R
Endogenous; X, Y
Exegenous; Z,N,P,R
How should enter the endogenous variable and exogenous variables in VAR?
If I enter the above exogenous and endogenous variables, Eviews gives the solution for all exogenous variables just thinking as a seperate and big equation for each X and Y. For example, Y is only depend on X,Z,M,N, and not depends on P,R. But eviews gives a solution of Y=f(X,Z,M,N,P,R)
Is it possible to get a system solution in VAR; like Y=f(X,Z,M,N) and X=f(Y,P,R)?
Thanks in advance
system of equations in VAR?
Moderators: EViews Gareth, EViews Moderator
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13319
- Joined: Tue Sep 16, 2008 5:38 pm
Re: system of equations in VAR?
The only way to estimate that is in a System object, I believe.
Follow us on Twitter @IHSEViews
Re: system of equations in VAR?
Dear Gareth,
Thanks for reply.
Could you please explain how will it be?
I made system object and entered two equations. But there is no VAR option in estimate part. There are least square, gmm etc but no VAR.
Thanks
Thanks for reply.
Could you please explain how will it be?
I made system object and entered two equations. But there is no VAR option in estimate part. There are least square, gmm etc but no VAR.
Thanks
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13319
- Joined: Tue Sep 16, 2008 5:38 pm
Re: system of equations in VAR?
a VAR is just a series of least squares equations, so you can estimate the system by least squares.
Follow us on Twitter @IHSEViews
Re: system of equations in VAR?
You need to add lags of the variables up to the VAR level you need.
y1= c(1) + c(11)*y1(-1) + c(12)*y1(-2) + ... + c(21)*x1(-1) + c(22)*x1(-2) + ... + ...
and so on and do the same for the second equation according to the structure you need!
An easier way is to use OLS (command: ls) on equation by equation, unless you assume the correlation b/n equations and want to model it!
y1= c(1) + c(11)*y1(-1) + c(12)*y1(-2) + ... + c(21)*x1(-1) + c(22)*x1(-2) + ... + ...
and so on and do the same for the second equation according to the structure you need!
An easier way is to use OLS (command: ls) on equation by equation, unless you assume the correlation b/n equations and want to model it!
Who is online
Users browsing this forum: No registered users and 33 guests