A Problem of No Stable Equilibrium

Hi, All

I have a question regarding the GMM estimation. When I run the mod file, I get a message as follows.

I don’t understand how the violation of the BK condition occurs. My model is an NK model of a small open economy with tradable and non-tradable goods. Two nominal variables are not stationary in my model: the non-tradable good price and the nominal exchange rate. Besides these, all the other variables are stationary. I ENCOUNTERED THIS ISSUE when I was trying to match the moments regarding stationary variables. Can you help me figure out how it can occur? I have tried to run the model using the calibrated parameters (excluding the estimation and moments-matching parts and just run the model under the basic parameter configuration, which works without violating the BK condition). Therefore, I feel confused and need some clarification.

Attached are the .mod file and the data.
database.xlsx (16.3 KB)
GMM_estimation.mod (3.3 KB)

Thank you!

By default, unit roots are counted as unstable roots in the method moment routines. Use

method_of_moments(mom_method=GMM, qz_criterium=1.000001,...

to count them as stable.
database.xlsx (16.3 KB)
GMM_estimation.mod (3.3 KB)

1 Like

Thank you, Prof. Pfeifer. That’s very helpful.

3 posts were split to a new topic: Bayesian estimation with unit roots