Why? Because calibrating to your own data removes potentially erroneous generic assumptions.
Consistency lets us refine and use appropriate modelling assumptions. These assumptions are based on fact and not guesswork.
... and effectively calibrate models.