Page 1 of 1

BK conditions

PostPosted: Thu Dec 19, 2013 10:00 am
by Dave
Dear all,

I receive the following "weird" error message:

"There are 8 eigenvalue(s) larger than 1 in modulus
for 8 forward-looking variable(s)

The rank condition is verified.

Error using print_info (line 46)
Blanchard Kahn conditions are not satisfied: indeterminacy due to rank failure".

I am curious under what condition Dynare will report this, saying the rank condition is verified, and at the same time complaining indeterminacy due to rank failure.

I suspect that one eigenvalue is larger than one but extremely small (1.001). Does it make sense. Thanks in advance.

Best,
Dave

Re: BK conditions

PostPosted: Thu Dec 19, 2013 12:23 pm
by jpfeifer
Try running model_diagnostics.

Re: BK conditions

PostPosted: Thu Dec 19, 2013 1:02 pm
by Dave
Dear jpfeifer,

Thank you for your reply.

I did run the test and no error message reported.

Best,
Dave

Re: BK conditions

PostPosted: Thu Dec 19, 2013 1:28 pm
by Dave
I have now attached the code.

Any comments and suggestinos are welcome! Thanks.

Best,
Dave

Re: BK conditions

PostPosted: Fri Dec 27, 2013 9:00 am
by jpfeifer
It could be a problem of numerical accuracy. Have you tried different parameterizations?

Re: BK conditions

PostPosted: Sun Dec 29, 2013 1:42 pm
by Dave
Dear Johannes,

Thank you for your suggestion. I also suspect the same. Most likely there are "dividing zero" situations in the computation process. The problem is that there is not that much I can play around with the parameters. I don't have a big hope resolving the problem. But, again, thanks.

Best,
Dave