Two country model eigenvalues problem

Hi, Professor! These days I am suffering from solving the dsge model. It always reports “One of the eigenvalues is close to 0/0 (the absolute value of numerator and denominator is smaller than 0.0000!”
The attached file includes my codes.

code0429.zip (7.8 KB)

Check this method:

\texttt{resid } ;
\texttt{steady } ;
\texttt{check(qz_zero_threshold=1e-20)} ;

Altough this problem may be a for a zero value for one of the steady states in your model, specifically when you have a ratio in your equations.

Thanks for your reply. I update the new version. According to your suggestion, it reports that: There are 21 eigenvalue(s) larger than 1 in modulus
for 22 forward-looking variable(s)

The rank condition ISN’T verified!

MODEL_DIAGNOSTICS: The Jacobian of the static model is singular
MODEL_DIAGNOSTICS: there is 2 colinear relationships between the variables and the equations

Check your predetermined variables timming in your model equations again.

Here, I assume T(t) = (1-delta(t))*T(t-1) + M(t) to capture the evolution of the trade relationship, this is an issue? I will recheck my codes. Thanks a lot.

In many cases Blanchard-Kahn failure is due to the timming issue for one of the endogenous variables.

As @eisamabodian correctly pointed out, this is most often a timing issue. The AR-process you outlined is not necessarily a problem. But given the size of your model, the only way out is usually to simplify the model considerable to see what causes the problem.

Thanks for your advice. I will recheck the setting

Hi, prof. I found a question about the steady state. In theory, if walras theory holds, it should give the same steady state result when ruling out a equation. However, in my symmetric two-country model, the results are not symmetric. I am curious about why this happened, since I have checked the coding written down in a symmetric way.
In my model, I introduce the search frictions between importers and exporters. Could you give me some clues to check what is wrong? Thanks again.

Hi prof. I assume the T(t) = (1-delta(t))*T(t-1)+M(t), where M(t) is new formed. delta(t) is an exogenous shock, I should make a change about the time notation?

  1. That is indeed unusual. Either there is still a mistake or the model has multiple steady states and the solver simply converges to different ones when altering the equations.
  2. The AR equation timing

looks correct.