System weighted R-squared

[paula_rossi]

I am aware that single equation R-squared measures are not appropriate to infer the goodness of fit of a system of equations since the R-square is only maximized with OLS whereas the purpose of a system estimation (such as 3sls) is to get accurate estimates of the coefficients.

Unfortunately, Eviews only reports single equation R-squares when estimating a system.

Is there any way to compute the system weighted R-squared or any other measure of goodness of fit of the whole system?

For your information I am using Eviews 5.1.

Thank you very much.

Paula