VAR exogenous variables

[lucaski]

I am trying to make a program that estimates a var, but I do not know how to make it estimate with exogenous variables. How can I do it?

[EViews Gareth]

As stated in the Object Reference:

Code: Select all

var var_name.ls(options) lag_pairs endog_list [@ exog_list]

Thus, it would be:

Code: Select all

var myvar.ls 1 3 y1 y2 y3 @ x1 x2

[pence5]

I have a related question, I ran this regression using an eviews program:

var myvar.ls 1 2 y1 y2 y3 @ x1(-1) x1(-2) x2(-1) x2(-2)

Then I opened the myvar object and then used the dialogue box called 'Estimate' to see what it was doing. All the variables were in the right boxes in the dialogue box. Then, without changing anything, I clicked ok (so it re-estimated the equation) and all the coefficients changed (by quite a lot!).

I have tried estimating a number of different VARs with exogenous variables and find I always get different results if I use the code above in a program compared to when I open the VAR object and re-estimate the VAR using the windows (and I am not changing anything).

Why is this? Is the program making different underlying assumptions to the VAR estimated using the dialogue boxes?

I am using Eviews 7.2.

Thank you in advance for your assistance.

[EViews Gareth]

The dialog adds a constant to the exogenous variables list. Your command doesn't have one.

[pence5]

Thank you very much for your quick and helpful reply!