Error when estimating dsge model

[MAIGA Issaka]

Hello everyone,
Can someone help me to handle this problem please,
"The process is saddle-path stable.
Warning: There may be a unit root in the model. Unit roots caused by nonstationary state variables are not allowed."

[Andrew Musau]

Are you supposed to check for stationarity of the variables prior to estimating DSGE models? I do not know as some of the best economists advise me not to bother learning about these models, and I listen. See, e.g., https://academic.oup.com/oxrep/artic...1-2/70/4781816. That aside, it may be the case that you need to check that the time-series are stationary and if not, difference the nonstationary series prior to estimation. See

help dfuller

for testing whether a variable contains a unit root.

[Mario Ferri]

When estimating a DSGE remember you are in steady state. Keep that in mind!

[MAIGA Issaka]

Thank you all for your reactions, I'd done dfuller test and differentiated the nonstationary variables but I still have this error message...

[Andrew Musau]

Nothing more to offer here. Look at the manual entry of the command to see if there are some suggestions. If no one else in the forum answers, consult with people who work with these models.

[MAIGA Issaka]

Noted with thanks Mr Andrew...