Instrumental variables

[pgrign]

Hi everybody,

I have a really simple system of equation in which I determine the term spread in the US and Germany.

D(GER10Y-EUR3M) = C(1) * D(GER10Y(-1)-EUR3M(-1)) + C(2) * D(USA10Y-USA3M) + C(3) * D(LOG(EXCH)) + C(6) * D(VIXEU)
D(USA10Y-USA3M) = C(11) * D(USA10Y(-1)-USA3M(-1)) + C(12) * D(GER10Y-EUR3M) + C(13) * D(LOG(EXCH)) + C(16) * D(VIX)

I tried to instrument those variables with the Central Bank rates, but I get an error saying "Near singular matrix". The problem is that, WHICHEVER variable I use, I ALWAYS get the same error.
Am I doing something wrong? Am I choosing the wrong instruments? (In that case do you have any suggestion?).

Maybe it's just a bug (I am still using Eviews 5)

Thanks!

[EViews Glenn]

Quick question. Do you get singularity if you estimate using OLS?

[pgrign]

Quick question. Do you get singularity if you estimate using OLS?

No

[EViews Glenn]

Sorry for the delay, I didn't see your response.

Hard to say.

Can you post the workfile? With a complete description of the instrument specification you used (ideally, with the system object saved in the workfile).