Yeah, pretty much all systems of governance ultimately evolve until their primary purpose is actually ensuring the survival of the system of governance and anything else it accomplishes is kind of a side effect. It's probably some sort of informational axiom of rules systems in general whether bureaucratic or biological or whatever.

Hell, DNA is just rules about what you can build and it's primary purpose is just making sure the rules survive. All the wonderful complexity and diversity of life is a side effect of the little changes necessary to propagate the rules.

Assigning single purpose to things is not necessary. "Systems are what they do" is a quote for a reason.

I think in addition to rules survival and admin self-concern, people genuinely underestimate how much maintenance and effort go into accomplishing goals in an organized, communicable, trustable way. It is also why AI is not as successful as people thought it was going to be at taking over jobs.

If you think the only value add to a business is the business output, you are taking admin work for granted.

In a way the bureaucracy takes on a life of its own. I think it’s only external pressures that’ll keep the bureaucracy in check, as in if the organization is at risk of dying the interests are aligned so that a more symbiotic relationship is necessary. When organizations are not at risk, either through massive initial success or state intervention (ZIRP) then feedback loop is cut and the bureaucracy will run rampant.

And that's why command economies fail. They fail in the same way that firms do, except that because the whole economy is one giant firm, you can't get the you need to remove entrenched bureaucrats until the situation gets so bad that you have a revolution or lose a war.