ordered probit & make model

[giuseppe]

good morning, I have just started using EViews 7 and I am trying to use an old code, that was working with EViews 4 until a few days ago, to iterate out of sample forecasting; it worked with EViews 4 but now I get an error. The code is:

for !horizon=1 to cont
equation eqopuno.ordered tre c esog1(-1) esog2(-1)
eqopuno.makemodel(modelopuno)
eqopuno.makelimits
modelopuno.solve(n=f)
....
next

What is strange is that the model doesn't work but if I create the equation from "quick" and then manually create the model it works, and even if I use a code to create an equation without specifying the name:
ordered tre c n1ortog(-1) tre(-1) eynorm(-1) rit1(-1) (vix(-1)-1)
then I can create the model manually, but if I estimate the equation with:
eqopuno.ordered tre c esog1(-1) esog2(-1)
then I can not create the model neither manually...

I hope someone can help me, thanks!

[EViews Gareth]

Hard to help without knowing what the error is.

[giuseppe]

"the model contains errors and could not be compiled in "do_modelopuno solve"

in the model I find: "err @identity is not a genr or series expression function in: ".eqopuno"

but if I create with the code an equation with the same time series (endogenous and exogenous) but with no name (ordered tre exogenous1(-1), exogenous2(-1)) and the create the model manually, it works and i geto no error messages, so the problem is not in the series but in the instruction wich gives the name to the equation it is estimating...

[EViews Gareth]

Looks like a bug in EViews, I'll investigate.

[giuseppe]

thanks a lot!

[EViews Gareth]

Latest patch should fix it for you.

[giuseppe]

it works with the new patch, thank you!!

[beatrice]

Latest patch should fix it for you.

Hi, I couldn't download the patch above. isn't it available anymore? thanks!

[EViews Gareth]

Try reloading that page.