Perfect forsight simulation fails with a new steady state

Hello all,
I’m working on an extended version of Gertler and Karadi (2011) with government debt. My problem is that the deteministric simulation of the model fails when I change the steady state. Initially, I solved the steady state by imposing two different values for two variables (the real rate on capital and that on bonds) and I managed to have both deterministic and stochastic simulations. Then, I wanted to impose the same value for the two variables at the steady state (i.e. for the two rates on capital and bonds), but in this case the perfect simulation fails even though the stochastic one is working. Do you have any ideas to solve the problem?
srff9.mod (9.4 KB)

@dben We try to not delete posts even if the problem has been solved as they may be useful for other users. What was the solution?

In order to get the deterministic simulation, each equation in the model should give one current endogenous variable (at time t). When I changed the steady state, the unique current variable in one of my equations has been canceled out. To solve the problem, I added 10^-10 to the imposed value in my new steady state, such that the current endogenous variable in the equation does not disapear.

I have some problems using the perfect foresight solver of Dynare. I’m working on a deterministic model with two intermediate goods and one final good.
Where G is the final good using a Cobb Douglas:
G_t=K_t^\alpha E_t^\gamma((1+g)^tAL_t)^{(1-\alpha-\gamma)}. And E (energy services), is aggregated with a CES E=(\omega e_{r,t}^\varphi+(1-\omega)e_{c,t}^\varphi)^{\frac{1}{\varphi}}. The exogenous variable is a tax on the intermediate good e_{c,t}. Dynare runs properly, and finds the steady state with the steady state model, however when I’m trying to implement a temporary shock (a little one), Dynare does not converge. Any suggestion to fix this problem?for_dyr.mod (3.8 KB)