EViews.com

Hi !

When we estimated a VAR(2) model, we got a result of estimation.

But, When we individually estimated the each equation.
Then, error message 'Near singular matrix error.Regressors may be perfectly collinear' was displayed.

Is this Bug?

Your data is perfectly multicollinear. Note the R^2 of 1.0 in the VAR. EViews just didn't catch it in the VAR most likely because computer arithmetic is imperfect.

It looks as though X = 10 + 0.99*X(-1) exactly (or as close to exact as a computer gets).

Hi

Thank you very much for useful reply.