During the mode finding stage, I get a [Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 2.171813e-16.] error in the console (repeatedly – so much so that I think it’s causing MATLAB to crash after running for a day).
I’ve tried to suppress these messages with:
warning off;
warning('off','MATLAB:singularMatrix');
warning('off','MATLAB:nearlySingularMatrix');
But these don’t seem to work. Is there a specific place in the mod file to place these commands?
1 Like
I keep running into the following error for an estimation exercise that does not involve the ZLB:
Piecewise linear Kalman filter: There was a problem in obtaining the likelihood
I’ve taken a look at the slides and code for the 2022 Advanced Workshop for some advice. But are there any additional example mod files for OccBin estimation?
1 Like
No, I am not aware of other examples. Maybe you could provide me with the mod-file to look into the error message. I guess it could be more informative.
PM’d.
Also, a note to MATLAB users: there seems to be an issue that Mathworks is aware of where some commands cause the Command Window to be totally blank even if code is executed successfully.
Unfortunately, estimation (with OccBin) is one such case on both my Intel iMac and my Apple Silicon MacBook Air. Both are running MATLAB 2024a with Dynare 6.1 but with x86 and ARM builds, respectively.
It makes debugging even more difficult.
@jpfeifer
After doing some testing, there might be some kind of issue with Dynare OccBin (ver 6.1) and MATLAB, which causes output in the MATLAB Command Window to not properly show for both OccBin simulations and estimation.
Your steady state file contains the line
varphi = 1/4; clc% 1/3 in ABK
The clc will clear the screen on every single call of the steady state file. That explains the missing output.
1 Like
Thanks @jpfeifer – a stupid typo on my end