One of the eigenvalues is close to 0/0 (diagnostics okay!)

Stochastic simulations
1

Dear all.

Running my attached mod file, I get the error “One of the eigenvalues is close to
0/0”. I checked the model diagnostics, which say that “no obvious problem is detected” and I’m quite sure that my steady state is correct. But if I reduce the qz_zero_threshold I get that “BK conditions are not satisfied”. Where can the error be?

Thanks a lot for every hint!
Best regards

bb.mod (4.4 KB)

2

Usually, there is a timing error involved. Equations like

Y(-1)            = c + K(+1) + B(+1) + L(+1) + r*L;

look strange. Why is output predetermined?

3

Thanks a lot Prof. Pfeifer! This equation for Y(-1) should reflect the green equation and the selected budget constraint. Output at time t is only available at t+1. But I‘m not sure if I‘ve implemented it correctly.

image
3EF1A94A-FD6B-423F-B052-86EE949E68D6.png1136×640 88.6 KB

4

This one is tricky to implement. So output really is predetermined in that equation. But you need to be careful when you combine equations. Equation (2) is a static one with K and A being predetermined states. Equation (3) cannot be entered this way into Dynare, because there would be expectations around the whole equation in this case.