Hi Michel,
I have been estimating a model with unit root shocks and measurement error in some stationary variables of the model.
The estimation (mode and mh) works fine, but the computation of smoothed shocks, smoothed observation errors and smoothed and historical variables seems buggy. I am enclosing a mod file of a simpler example that replicates the problem. I guess you might be already aware of this, but I was wondering whether it had been fixed.
The mod file is based on your example file mod1.mod in the Dynare_v4 documentation. I have simply added an extra (stationary) variable to the model with measurement error in it.
Thanks
Matteo