Endogenous Growth Component and Steady State problem

Dear Dynare People,

This is the first time I post in the Forum, I have a DSGE model that extends CET (2016) Labor Force Participation model with some endogenous growth component.

I solved for the steady states, and check for the residuals, I can see residuals can be very close to zero for all equations.

If I use steady(solve_algo=3,maxit=10000);

Dynare Told me rank condition isn’t verified.

If I use steady(solve_algo=4,maxit=10000);

Dynare would say it is not possible to find steady states.

I also find previous posts saying that it might not be possible to have steady states for endogenous growth model.

At this stage I am not sure where the problem really comes from. Is it because wrong computation, steady states are infinity, or there is something wrong in the equations, or it is a general feature of endogenous model.

I would really appreciate any suggestion or advice, or references to look at.

Thanks so much for your time to look at my problem!!simple2.mod (9.0 KB)

That is hard to tell. If solve_algo=3 finds a steady state, then it can be endogenously computed. The problem is that there may be multiple steady states. So you should have a look at the steady state values to see whether they make sense in both sign and (relative) size.

Hi professor,

Thanks for answering my question.
I tried a much simpler RBC model with only endogenous growth block, somehow when I change the timing of inflation in the FOC to bond one period before, the model runs.
I also changed only the timing of inflaiton in previous mod file for 35 eqns, it also run with solv_algo=3.
It is kind of puzzling for me, because I am doing something counter-intuitive but it works.

For steady states, I did compute in Matlab and make sure they are right.

Just a general question, if we start with any model, suppose check for steady states and everything, is it possible just because of parameterization, Dynare can fail?

Is there any other way to run the model? Or it is always the set up does not make sense, what else can we do?

Thanks a lot!

simple3.mod (3.9 KB)

It can come from parameter values. The Taylor principle is one example of this. But generally, a timing issue is more likely. What you write also suggests a timing issue.