Yeah, it's weird. Early economists clearly didn't have the data or computing capability available to do analysis on the real world. So instead the took some inspiration from it and then built these simplified models, that were at least possible to study and make conclusions about. However, because real-world economic outcomes are so important, politically and socially, people ended up taking their conclusions and trying to apply them back to the real world.

It's as if, after early AI researchers who took some inspiration from biology to create neural networks, the medical establishment then took their results and decided to use them as a guide to performing brain surgery.