EViews.com

Hi,
I am encountered a problem regarding testing significance of estimated coefficients.
I have estimated the following two equations independently using Eviews:

Y1=a0 + a1*X1 + a2*Y2 + e
Y2 = b0 + b1*X2 + b2*Z1 + I

e and i are respective residuals. I hypothesize that Z1 affects Y2 and Y2 affects Y1. I am now interested to test significance of a2*b2=0. I have estimated these equations independently using GMM to tackle endogeniety and have not estimated them as a system. I need help to know how can I test its significance. It is of critical importance for my study. Kindly, guide me how can I proceed. Your valuable comments will be highly appreciated.

Regards

Hi,
I am encountered a problem regarding testing significance of estimated coefficients.
I have estimated the following two equations independently using Eviews:

Y1=a0 + a1*X1 + a2*Y2 + e
Y2 = b0 + b1*X2 + b2*Z1 + I

e and i are respective residuals. I hypothesize that Z1 affects Y2 and Y2 affects Y1. I am now interested to test significance of a2*b2=0. I have estimated these equations independently using GMM to tackle endogeniety and have not estimated them as a system. I need help to know how can I test its significance. It is of critical importance for my study. Kindly, guide me how can I proceed. Your valuable comments will be highly appreciated.

Regards

As a practical matter, the easy way to do the test is to estimate the two equations as a system. The problem is you need to account for the covariances between the coefficients, which in turn depends on the covariance between the errors in the two equations. System GMM will take care of this for you.