I rewrite one well-worked model into a smaller one by reducing some variables and equations, i.e., change some endogenous variables into parameters and calibrate them to zero. But after this, MODEL_DIAGNOSTICS shows that there are 6 colinear relationships in my model and almost all equations are covered (each colinear relationship lists more than 50 equations, while there are about 60 equations all together in my model). This confuses me a lot and I can’t find the reason. Has anyone once run into such a situation? Thanks in advance.
Without knowing the model it is impossible to tell. Is there just collinearity or is there a unit root driving this?
Thanks for replying.
I am not very sure about the unit root problem. But there is no eigenvalue exactly equalling to 1, does this mean there is no unit root in the model? Btw, one more thing might cause this is that not only did I reduce some variables but also make some changes to the calibration methods. However, I doubt if this can be the ture reason.
In this case, I can only recommend simplifying the model to see where the issue appears.