Hi,
I'm working on a large model which contains more than 4000 equations. Solving this model both under EVIEWS 7.2 and version 6 (standard edition, Feb 2011 build) leads to different results for the endogenous variables, even though it's the same workfile and the solution settings are the same as well (dynamic, deterministic, Gauss-Seidel ). This is really puzzling…
What could be the cause of this? Thank you in advance for your assistance.
Best regards
Seb
Different EViews versions -> Different model solutions?
Moderators: EViews Gareth, EViews Moderator
-
- Non-normality and collinearity are NOT problems!
- Posts: 3775
- Joined: Wed Sep 17, 2008 2:25 pm
Re: Different EViews versions -> Different model solutions?
Check that the starting values of the parameters are the same in both cases.
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13319
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Different EViews versions -> Different model solutions?
Any chance you could post the workfile?
Follow us on Twitter @IHSEViews
Re: Different EViews versions -> Different model solutions?
Thank you for replying so quickly. Do you mean the starting values in the C vector? Yes, they are the same in both cases.
And I selected "actuals" to be used as "preferred solution starting values" in the iterative solution procedure. Are there any other important starting values?
Perhaps I should try to explain my problem in more detail. I used EViews 7.2 (at the office) to estimate the equations, to create a model and to add the equations to the model. Afterwards I saved the workfile. At home I only have EViews 6. but my guess was that these two versions should be quite similar. So I took the saved workfile and solved the model under EViews 6 at home again. Surprisingly, the baseline solutions are not equal although I just opened the saved workfile and solved the existing model. I didn't re-estimate any equations or changed any settings. Could different default settings cause my problem?
Many thanks for your help
Seb
@gareth: I tried to send you the workfile along with a pm, but the file size is too large (6 mb zip-file). Any idea?
And I selected "actuals" to be used as "preferred solution starting values" in the iterative solution procedure. Are there any other important starting values?
Perhaps I should try to explain my problem in more detail. I used EViews 7.2 (at the office) to estimate the equations, to create a model and to add the equations to the model. Afterwards I saved the workfile. At home I only have EViews 6. but my guess was that these two versions should be quite similar. So I took the saved workfile and solved the model under EViews 6 at home again. Surprisingly, the baseline solutions are not equal although I just opened the saved workfile and solved the existing model. I didn't re-estimate any equations or changed any settings. Could different default settings cause my problem?
Many thanks for your help
Seb
@gareth: I tried to send you the workfile along with a pm, but the file size is too large (6 mb zip-file). Any idea?
-
- Non-normality and collinearity are NOT problems!
- Posts: 3775
- Joined: Wed Sep 17, 2008 2:25 pm
Re: Different EViews versions -> Different model solutions?
I suspect you need Gareth to come to the rescue...
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13319
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Different EViews versions -> Different model solutions?
email it to support@eviews.com, along with instructions on how to replicate what you're seeing in both versions.
Follow us on Twitter @IHSEViews
Who is online
Users browsing this forum: No registered users and 24 guests