EViews.com

Posted: **Mon Mar 31, 2014 11:27 am**

Hello,

I am new to Eviews7 and doing regression on the model by Heston and Rouwenhorst (1994):
R = α + β(i1) + β(i2) + γ(c1) + γ(c2) + e
where i represents dummies for industry and c represents dummies for country.

Since there is a problem of perfect multicollinearity, we have to impose the constraints that:
sum of all weighted beta = 0 ; sum of all weighted gamma = 0

How can I impose these two restrictions in Eview7 when I do the estimation? Please help and thanks!

Regards,
Learner

Posted: **Mon Mar 31, 2014 11:39 am**

Hello,

I am new to Eviews7 and doing regression on the model by Heston and Rouwenhorst (1994):
R = α + β(i1) + β(i2) + γ(c1) + γ(c2) + e
where i represents dummies for industry and c represents dummies for country.

Since there is a problem of perfect multicollinearity, we have to impose the constraints that:
sum of all weighted beta = 0 ; sum of all weighted gamma = 0

How can I impose these two restrictions in Eview7 when I do the estimation? Please help and thanks!

Regards,
Learner

Something like

Code: Select all

ls r = c(1) +c(2)*i1 + c(3)*i2 +c(4)*c1 + (c(2)+c(3)-c(4))*c2

Posted: **Tue Apr 01, 2014 3:31 am**

Thanks

I have just tried to input the following command:

Code: Select all

ls return c cs-(i2/i1)*cd e-(i3/i1)*cd f-(c4/c1)*cd hc-(i5/i1)*cd i-(i6/i1)*cd it-(i7/i1)*cd m-(i8/i1)*cd o-(i9/i1)*cd t-(i10/i1)*cd u-(i11/i1)*cd bz-(c2/c1)*au cn-(c3/c1)*au ch-(c4/c1)*au gr-(c5/c1)*au hk-(c6/c1)*au in-(c7/c1)*au jp-(c8/c1)*au ks-(c9/c1)*au uk-(c10/c1)*au us-(c11/c1)*au

but it keeps showing an error message of "Division by zero"

how could I solve it? Many thanks!

Posted: **Tue Apr 01, 2014 8:32 am**

type c=0.1 in the command window before estimating.

EViews.com

How to impose the following constraint?

How to impose the following constraint?

Re: How to impose the following constraint?

Re: How to impose the following constraint?

Re: How to impose the following constraint?