f-statistic in 3SLS

[22023661]

Hi,
what can I do to make eviews show me the f-statistic in 3SLS?
thanks

[EViews Glenn]

That's not an easy one to get since it requires estimation with a fixed weighting matrix. To be honest, you are best off simply doing the equivalent Wald statistic rather than trying to do the F.

[startz]

That's not an easy one to get since it requires estimation with a fixed weighting matrix. To be honest, you are best off simply doing the equivalent Wald statistic rather than trying to do the F.

The equivalent Wald statistic is the F-, no?

[EViews Glenn]

I was answering in the context of the question which was how to get the F-statistic (which I assumed, perhaps incorrectly, was the likelihood ratio) to appear. Since we don't have an F-statistic view, I was pointing the user to the Wald test view which will offer an asymptotically equivalent test statistic.

Now starz is correct in noting that there are equivalences that mean that the Wald is numerically equal to the LR in some settings. Perhaps in this one.

But I wasn't certain that this was a linear specification (and since I never remember all of the conditions for equivalence anyway), I opted merely to point the user to the asymptotically equivalent test (which may be numerically equivalent in the current setting). In any event, startz's comment is nonetheless a useful one.