What I find odd, is often the biggest proponents of the free market - with it's decentralised decision making process ( and redundancy of effort ), decide that total centralized, top down control is the best way to run their own company - as they think top down decisions and minimising waste through duplication is the way to go.
That's a slightly different angle - if I understand correctly it's about why firms exist at all - and if I were to summarise it's because lack of trust costs - a firms boundaries are created to balance transactional costs ( within the firm it's low, between firms or individuals it can dominate ) versus the costs of perhaps not being market efficient.
My argument above is about not what shapes a firm's boundary but how it operates internally - too much top down control potentially risks exacerbating the risks associated with being a company and also potentially increases internal transactional costs as well - worse of both worlds! - all that ceremony around decision making, time spent justifying existences, inability to just act.
Obviously as I said above - it's a balance - just as it is with country/international systems.
> In a “free market” you can choose what firms you work for and with, except the government.
Eh? Nobody ever leaves one country for another?
> The government would be more efficient in an autocratic leadership. But government efficiency is not the societies efficiency and well being
I think the free market proponents would say otherwise - one key problem is the myth of perfect information - you are imagining it's possible to concentrate all the required information to make any correct decision into a very small group of people at the top ( and assuming these people are competent and not corrupt). One of the ideas behind why distributed markets work is the information about what is needed is communicated by the mechanisms of the market itself.
And to bring it back to the original post - does the CEO have all the information required to direct others to meet the customers need across all areas - or is it better to use the collective intelligence of the entire organisation?
And in the real world, not even this is true, due to the cost of satisfying OKRs--opportunity cost is one kind of cost, but optimizing an OKR can of course negatively impact other valuable things that aren't represented by OKRs.
Finding and fixing product problems requires decentralized decision making and trust.