Theoretical moments are NaNs

Hi, I have simulated model (attached) and received NaNs for theoretical moments of all variables but two. For simulation of empirical moments, it works. Is there any bad specification of the model? Should I worry about it?
Thank you for answer. Miroslav
housing_open6.mod (14 KB)

Your model has a unit root. In this case, the moments to not exist. This is the reason Dynare dispays them as NaN.

Dear professor. So, how can I solve this problem? And Dynare reports the impulse response plot, but there is a lot of NaN in the variance decomposition. And, my model has a unit root, because of random walk, but my research needs the variance decomposition. How can I solve this problem~

You cannot decompose an object that is infinite. That is the reason why the unconditional variance decomposition does have NaN.

Dear, professor. If I consider a shock process as shown in the figure, can I solve this unit root problem?

Dear professor, I found that the variance decomposition sum of some variables is not equal to 1, why is that

  1. As long as |\rho|<1, the exogenous process will not introduce a unit root.
  2. For the simulated variance decomposition, the explanation is printed on the screen. If this happens for the theoretical decomposition, I would need to see the codes.

Please read my answers carefully.

If you specify periods>0, then you get

Note: numbers do not add up to 100 due to non-zero correlation of simulated shocks in small samples

There is no way to change this.

Thank you sir!