causal
Calibration can fit data without approaching reality
Calibration—taking a purely theoretical model, feeding in shocks and adjusting them until the model replicates observed data—creates a false sense of validity, because enough tinkering will eventually hit the mark without bringing the model closer to reality, and it absorbs effort into programming rather than thinking about the real world.
causalpending
Speaker
Guillermo CalvoEvidence Quote
“calibration, what you do is have a model which has become a completely theoretical construct; put some shocks into the model, play with the shocks, and then see if the model replicates what you see out there. Well, if you keep playing with the shocks, eventually you hit the mark. But that doesn't mean you are closer to reality.”
Created: 6/13/2026, 3:35:48 AM
Your notes
Loading notes...