Hello, Professor @jpfeifer.
I’ve noticed that many researchers – especially those not directly involved in DSGE modeling – tend to be quite skeptical about the practice of parameter calibration.
Could you please explain what arguments can be made in favor of calibration, and why it is considered acceptable or appropriate in the context of DSGE models?
It depends on what the research question is. You can make theoretical points without perfectly tailoring a model to a particular set of data at a particular place and time. But that does not allow you to, e.g., quantify uncertainty.
2 Likes