> I'm sure economists/data-scientists have models for that, and if there would be any obvious sign, we would know it by now.

There certainly are people with models. The problem is there are way more occurrences of people using some model to say a crash is imminent than there are crashes.