When I attempt to run Dynare and use the identification command, Matlab completely crashes: it disappears without any trace from my screen. Is this routine not yet operational. Is there anyway to use it?
Thank you for any help and advice you can provide.
Rob Luginbuhl
Dynare forums
Forums for asking questions, posting comments and uploading examples related to Dynare
Does identification work in 4.2?
Forum rules
This forum is closed. You can read the posts but cannot write. We have migrated the forum to a new location (https://forum.dynare.org) where you will have to reset your password.
This forum is closed. You can read the posts but cannot write. We have migrated the forum to a new location (https://forum.dynare.org) where you will have to reset your password.
7 posts
• Page 1 of 1
Does identification work in 4.2?
by reluginbuhl » Thu Apr 07, 2011 3:39 pm
- reluginbuhl
- Posts: 53
- Joined: Wed Mar 04, 2009 2:41 pm
Re: Does identification work in 4.2?
by HoutanBastani » Fri Apr 08, 2011 1:22 pm
Please post your .mod file.
Best,
Houtan
Houtan
- HoutanBastani
- Posts: 197
- Joined: Fri Jan 22, 2010 4:11 pm
- Location: Paris, France
Re: Does identification work in 4.2?
by reluginbuhl » Fri Apr 08, 2011 2:48 pm
I actually first noticed this problem using a larger model. But I just checked, and I get the same thing using the RBC model. Here are the files for the RBC model.
- Attachments
-
RBC_ID_steadystate.m- steady state m file
- (1.89 KiB) Downloaded 61 times
-
- RBC_ID.mod
- mod file for RBC
- (1.41 KiB) Downloaded 87 times
- reluginbuhl
- Posts: 53
- Joined: Wed Mar 04, 2009 2:41 pm
Re: Does identification work in 4.2?
by rattoma » Tue Apr 12, 2011 8:59 am
Hi,
there were a number of problems in your mod file
1) you observe a non-stationary variable: identification uses theoretical moments of observed variables, and variance is infinite for non-stationary ones;
2) Y_obs was defined in such a way that it does not have a static solution (dA is not 1 in steady state)
3) observation trend for Y_obs is not log-linear
I attach a new project where you observe dY_obs. Then, the whole list of parameters is not identified, and I have re-set the list of params to an identifiable one.
If the project does not work with the official 4.2.0, there are debugs in the 4.2 branch on the official dynare repository.
best
Marco
there were a number of problems in your mod file
1) you observe a non-stationary variable: identification uses theoretical moments of observed variables, and variance is infinite for non-stationary ones;
2) Y_obs was defined in such a way that it does not have a static solution (dA is not 1 in steady state)
3) observation trend for Y_obs is not log-linear
I attach a new project where you observe dY_obs. Then, the whole list of parameters is not identified, and I have re-set the list of params to an identifiable one.
If the project does not work with the official 4.2.0, there are debugs in the 4.2 branch on the official dynare repository.
best
Marco
- Attachments
-
- RBC_ID_steadystate.m
- (1.9 KiB) Downloaded 274 times
-
- RBC_ID.mod
- (1.49 KiB) Downloaded 99 times
- rattoma
- Posts: 51
- Joined: Tue Jan 04, 2005 10:48 am
Re: Does identification work in 4.2?
by reluginbuhl » Wed Apr 13, 2011 2:49 pm
Marco,
Thank you very much for your help. Your comments are concise, clear, and helpful.
Your corrections and improvements unfortunately do not result in an mod-file that will run in Dynare 4.2. Your mod-file does run in 4.1.3 if I use the Matlab routine 'myboxplot.m' from either 4.2 or 4.1.2 and add the Matlab routine myprctilecol.m (found on the Dynare website and cut and paste into an m-file).
In 4.2 the identification command causes Matlab to crash. This is due to a problem with the Dynare file Dynare\mex\matlab\win64-7.5-7.7\mjdgges.mexw64. Does anyone know if this is a problem with Dynare 4.2? Or is it something to do with my Dynare installation or version of Matlab (7.7.0.471 R2008b)?
Thank you for the help.
Sincerely,
Rob Luginbuhl
Thank you very much for your help. Your comments are concise, clear, and helpful.
Your corrections and improvements unfortunately do not result in an mod-file that will run in Dynare 4.2. Your mod-file does run in 4.1.3 if I use the Matlab routine 'myboxplot.m' from either 4.2 or 4.1.2 and add the Matlab routine myprctilecol.m (found on the Dynare website and cut and paste into an m-file).
In 4.2 the identification command causes Matlab to crash. This is due to a problem with the Dynare file Dynare\mex\matlab\win64-7.5-7.7\mjdgges.mexw64. Does anyone know if this is a problem with Dynare 4.2? Or is it something to do with my Dynare installation or version of Matlab (7.7.0.471 R2008b)?
Thank you for the help.
Sincerely,
Rob Luginbuhl
- reluginbuhl
- Posts: 53
- Joined: Wed Mar 04, 2009 2:41 pm
Re: Does identification work in 4.2?
by rattoma » Wed Apr 13, 2011 7:01 pm
Hi,
I think you have either to use 4.1.3 or build the debugged pre-processor for 4.2 dowloading the source codes from the dynare repository (it will be released as dynare 4.2.1).
best
Marco
I think you have either to use 4.1.3 or build the debugged pre-processor for 4.2 dowloading the source codes from the dynare repository (it will be released as dynare 4.2.1).
best
Marco
- rattoma
- Posts: 51
- Joined: Tue Jan 04, 2005 10:48 am
Re: Does identification work in 4.2?
by reluginbuhl » Thu Apr 14, 2011 3:25 pm
Marco,
Thank you. I really appreciate this information.
-Rob
Thank you. I really appreciate this information.
-Rob
- reluginbuhl
- Posts: 53
- Joined: Wed Mar 04, 2009 2:41 pm
7 posts
• Page 1 of 1
Who is online
Users browsing this forum: Google [Bot] and 8 guests