Identification_rank of J is deficient

[sadia]

Dear Johannes,

I have read several posts related to this issue. But I am confused what are the sources of identification problem in my model. I have attached the mod file, data, some part of log file and others. Its always the rank J deficient. Its a stationary model. data are hp filtered. The model still runs. What should I do to fix this? Is it because of the priors? Many thanks.

Kind Regards.

[jpfeifer]

1. A general remark: You are not supposed to use a two-sided HP-filter for likelihood-based estimation
2. I cannot replicate the message with the unstable version (Dynare 4.5). I guess this was a numerical problem that has been fixed.

[sadia]

Hi Johannes,

1. I am using one-sided hp filtered data. I can try using linear-detrending instead if you suspect filtering can be a reason.

2. I am using Dynare version 4.4.3. I never used unstable 4.5. The file I sent you still produce the following message if I use the simple identification command. I have checked it once again. I don't understand why its happening, estimation runs smoothly.

==== Identification analysis ====

Testing prior mean
Evaluating simulated moment uncertainty ... please wait
Doing 378 replicas of length 300 periods.
Simulated moment uncertainty ... done!

All parameters are identified in the model (rank of H).

WARNING !!!
The rank of J (moments) is deficient!

Monte Carlo Testing

Testing MC sample

All parameters are identified in the model (rank of H).

WARNING !!!
The rank of J (moments) is deficient for 214 out of 250 MC runs!

Kind Regards,

Sadia

[jpfeifer]

1. Then it should be fine. Note that the data treatment has no bearing on the identification command. Identification is a theoretical concept here. Your data treatment can therefore not be the reason.
2. Then please try the unstable version (http://www.dynare.org/download/dynare-unstable). There have been several bugfixes around identification (http://www.dynare.org/DynareWiki/KnownBugs). This might explain why I don't get an error. From what I can see, your model is fine.

[sadia]

Hi Johannes,
Yes, you are right. It was a bug. I got the unstable version and it detects no problem relating to identification.
Thank you very much.

Cheers,

Sadia