Take one of the most widely-known principles in macro: the Philips curve. In its most basic form, it says that as unemployment decreases, employment increases.

Take one of the most widely-known principles in macro: the Philips curve. In its most basic form, it says that as unemployment decreases, employment increases. The trouble is, it doesn't work very well. So economists augment it with other unobservable variables like 'inflation expectations'. But if you go looking for evidence that these newer models are any better, 99% of what you'll find is theory. The rules of the game reward you for making elegant models, not for making good empirical predictions.