Dear porfessor,
This is a note from the picture
You mentioned that the model’s spread is based on the Kalman smoother, which seems to imply that you suggested there are multiple ways to compare the model to actual data. Am I correct?
Is there any literature or paper that discusses comparing model variables to actual data, which I can refer to?
Thanks
Oliver
The question is whether you compare ergodic properties like moments from the model to the data or whether you proceed with full information techniques that allow comparing data points. The latter is usually done in the context of estimated models.
Dear professor
I have already compared the model moments. However, in case I need to proceed with full information techniques in the future, I now want to learn how to compare the model with data in the latter case.
This is the estimation command and the Bayesian estimation results are stored in oo_. But I am not sure which the data I need.
I hope I’m not wrong about the way to compare data points.
Thanks again
Oliver
It’s all documented in the manual when you set the smoother option: Results are stored in oo_.SmoothedVariables, oo_.SmoothedShocks, and oo_.SmoothedMeasurementErrors.
For starters, I would suggest to skip estimation and use the calib_smoother command.
Dear professor
Your instructions have been extremely helpful and greatly appreciated.
Best
Oliver
