Dynare forums

i write a dynare code, it runs well in dynare4.0.4,however when using dynare 4.3.2, it shows errors with "no stable euquilibrium".I do not know why it hanppens. Is dynare4.3.2 request more strict conditions of the model? could anyone help me with this?
and another question:can dynare4.0.4 make "shock_decomposition" of the variables?
thanks

Could you post the mod-file?

Generally, when you have a mod-file that does not run in a previous version, there are issues in the model that make it borderline. There are models that have errors, but due to numerical rounding on computers don't show up in previous versions. I would thus rely on the more recent version.

The only thing I can think of that has been significantly changed from earlier version regarding the Blanchard-Kahn conditions is the treatment of unit roots in estimation compared to stoch_simul.

The reason is that your model has two unit roots. In this case, you must use

Code: Select all

lik_init=2

or

Code: Select all

diffuse_filter

to get the same behavior as in older versions of Dynare.