I am simulating a version of Bernanke, Gertler and Ghilchrist,1999 model on the housing sector. The model runs and generates IRFs but the endogenous variables are constant or non stationary. Any help to fix this would be highly appreciated.

Check the IRFs. If after, say, 1000 periods the IRFs do not return for the variables, they are non-stationary. If this is the case, find out why you have a unit root.

dear professor，

I had the same problem.Dynare hint I that “All endogenous are constant or non stationary”.May I ask whether this problem is caused by my wrong model setting?

Without seeing the model it is impossible to tell

There are two issues here:

- You are only having one shock with non-zero variance, which does not affect the variables you are interested in.
- All of the variables you are interested in are affected by a unit root. That implies they are non-stationary.

Thank you very much for your constructive help