IRFs plotted in spite of collinearity

[fabiac]

Hi there, my model runs without any apparent problem and I can even get IRFs. However, when I command "model_diagnostics", I get the error message saying that there are a number of collinear equations. What's the trick here? Should I trust the output of "model_diagnostics" even in this case? thanks a lot

[jpfeifer]

I have totally misspecified models seen produce sensible looking IRFs. Hence, I would be careful and look for a potential problem. However, there are usually also other warning signs like a close to singularity warning. Also, if your model contains unit roots, you might get collinearity warnings although everything is OK.