Certainly not physics - but shouldn't it be more like biology than economics? You can do experiments in biology (and with LLMs), the problem is you can't quite get to a cause-effect. The situation is just so complex. In both cases we are desperate to get an "explanation" that fits in a human brain and makes us feel like we understand, but it's out of reach.