I haven’t watched the clip, but I’m interested in the general topic, and can share an opinion about it. In my experience and education in organizational development, I’d say this is a truth, but not the only one.

In a company, it can make sense to oscillate between periods of centralization and decentralization of power and agency. There are economies and diseconomies of both.

When decisions are complex, highly various, often unique and non-routine, decentralized decision makes sense. But coordination of action doesn’t scale up that way. It’s hard to do something like build a large hydroelectric dam or high-speed rail system using decentralized power.

Think of a hospital. It has both an administration and a medical faculty. It is far more effective for medical decisions in the emergency room to be made by qualified people on the spot. It is also far more efficient for central administrators to figure out how much soap the hospital goes through per month, and where to get the needed product for the best price.

There is no efficient way to get information about all the craziness that comes through the emergency room door to a single central official to make medical decisions. There is also no efficient way to get information about aggregate soap use over time to rotating shifts of medical staff. One set of decisions should be centralized, another decentralized.

There’s not always an absolute answer to the degree of centralization that makes sense. Think of the difference between the iOS technological ecosystem and the Android ecosystem. One is much more centrally controlled and limiting than the other. One is much more open and diverse, but that results in a less uniform platform to develop for.

This illustrates that in some circumstances, both options are valid and effective, but there are tradeoffs associated with whatever option you choose.

Speaking of iOS vs Android, platforms illustrate yet another relationship between central and decentralized decision making - that of enablement.

Central control and creation of the platform (or standard) enables more local decentralized decision make by people using it. The internet itself is an example of this. Committees and consortiums define the protocols. Companies standardize technology architectures. Other companies create software platforms that run on this infrastructure. On top of all that infrastructure, an industry of app makers is enabled, and we can use those apps to gather information and make decisions locally.

A nation-state can be usefully analyzed as a platform, at least in part.

The right balance between centralized and decentralized control is also dynamic. It changes in time, which is yet another way the two forms of decision making are interrelated.

On a normal day, you can move around a city as you want. It’s better for everyone if you just make those decisions for yourself.

If a hurricane is barreling down on that city, it’s better if people obey the directives of people who are centrally positioned, getting information funneled to them.

Whenever large scale coordinated responses within tight timeframes is needed, the concentration of decision making authority is more efficient and effective.

But by that very same logic, blunders at this level can rapidly wreck things at a very large scale too, especially when conditions are different across the many sites impacted by the blunder - and those sites have all the time and resources they need to govern themselves.

The superiority of decentralized decision making authority and responsibility is real, in many situations.

Also, in many situations, it makes more sense for local collectives to take their direction from more central authorities.

And it can be good for some systems to oscillate in cycles between the devolution of power from the centre to the edges, and then to recentralize, and to repeat that cycle.

When conditions are such that central decisions are failing because local knowledge is needed to solve problems, devolving powers makes sense. But then the organization does some learning about regularities in those local problems, across many locations. That learning is ripe for routinization, but sometimes only a centrally positioned official can see the pattern across multiple sites.

Simultaneously, diseconomies of decentralization can build up, with resources squandered reinventing the wheel across locations, and different locations working at cross purposes, neutralizing or cannibalizing others’ gains, and generally become an unmanageable “herd of cats” that makes even local coordination difficult.

At those times, it can make sense to centralize things again, eliminate duplication of effort, reinforce a shared sense of direction, and standardize and routinize some of the learning discovered at the edges of the organization when things were more decentralized.

Then as the context changes, those routines and procedures can lose their relevance, and it can make sense to decentralize again, for another cycle of learning and local creativity.

When you are hired by a troubled organization to help straighten out a problem, all of this can be important to consider. I’m not a politician, but I can’t help but think these ideas function in politics as well.