Weird IRFs

[mshtm]

Hi all,
I have an extremely simple version of my model that I am trying to solve using dynare. When I input the equations in the standard way, I get an error for the Blanchard Kahn conditions not being satisfied (The model_diagnostics command doesn't reveal anything). However, when I lag the entry condition(which is like an Euler equation in my model) by one period, I do get a solution and get some IRFs. However the IRFs are very weird- they cycle multiple times even though they eventually converge back to the steady state. I have attached both versions of my mod file- in the first BK conditions don't hold and the second mod file attached generates funny IRFs. The steady states are correct as I used steady_state_model in dynare and double checked using fsolve in matlab. The variable N should be the only predetermined variable in the model.

I should add that I also tried inputting my model in levels (without exp) and also as a linear model after log-linearizing it by hand. But I get similar issues in all cases (in all cases, the model solves only if i lag the entry condition and generates very weird IRFs). However, I got the expected IRFs when I used the log-linearized version of my model in the Uhlig toolkit. Thus, I am very confused and can't figure out what I am doing wrong in dynare. I would greatly appreciate any help.

[jpfeifer]

This means you got some timing wrong and at least one variable does not conform to Dynare's timing convention of predetermined variables being dated time t-1. The IRFs you get when lagging the entry condition follow from a complex eigenvalue you introduced. However, if the Uhlig code works, it should be straightforward to translate it to Dynare with the correct timing.