I think you have one big piece of it: economic progress has a lot of search problems and it is impossible to master-plan it; consequently free intelligence beats centralized regulation. It's a bit out-dated now[0] but The Fifth Discipline distinguishes between 'detail complexity' (things that have a lot of bits you have to figure out) and 'dynamic complexity' (systems that have feedback loops and adaptive participants). It might simply be that handling systems with dynamic complexity is out of the reach of most humans. Economic regulation strikes me as something that can be particularly like a thing that modifies a dynamic system.

In fact, creating good policy in a modern economy might be so dynamically complex that no mind alive today can simultaneously comprehend an adaptive solution and act in such a way as to bring it about.

Perhaps, given this, we are simply spoiled by the effectiveness of certain economic actors (e.g. the Federal Reserve) in maintaining an monetary thermostat. Their success is not the norm so much as it is extraordinary.

0: which is humorous given this, because the Seinfeld Isn't Funny effect applies to things that become mainstream - insight and humor both disappear as the spark or joke become common knowledge