Is centralisation really more efficient?

This is a work in progress, as I’m collecting examples and data. Follow along by subscribing. Send me examples and arguments via email or twitter.

Businesses tend to grow as they add more areas of action to their structure. They do this because of economics of scale. And because we tend to believe that centralising efforts is a way to make them more efficient.

Governments tend to grow as they take on more and more regulations to oversee. They do this because regulations require inspectors, and arbiters, of all the edge cases that pop up over time. And because we tend to believe that centralising efforts is a way to make them more efficient.

This is such a pervasive idea that it’s hard to even describe where it starts. In most human collaboration, it seems to be a common truth that centralising efforts lead to more efficiency. My hypothesis is that this might not be true.

My idea

My hypothesis is that it is not efficiency, but control, that increases with centralisation. And we simply conflate the two, because it seems like a natural combination. Whenever we’re collaborating in small groups, like making dinner together, it’s simply a mess until someone takes the role as the chef.

I believe that works because a central authority can determine what the other actors should do by dictating the intended outcome. This is not efficient however. Because translating the central authority’s intended outcome to the actors comes at the cost of a lot of trial and error. And more interestingly: people trying to game the rules.

This is command and control. Not efficiency.

What is efficient and how it works is beyond the scope of this work. However I believe Yuval Harari is not far off when he speaks of stories as the thing that allows humans to collaborate in large groups. However I think it’s not any stories. But rather knowledge, an understanding of how the group does something, and a shared goal, an understanding of what the group is trying to achieve, that create effective collaboration.

What I’m doing to test the idea

So how can we test this idea? I can’t currently think of an experiment that would show this clearly. But I suspect that if I look for examples of organisation centralising to solve problems, I will not find any efficiency increases, only an increase of control.

This post will grow as I add examples. Please help out by sending me examples of organisations, large and small, that are centralising. You can reach me on email and on twitter.

Examples of centralisation left to check

The EU - not explored

Google M&A

Apple M&A - should be an interesting example, since their strategy is vertical integration of the production chain.

Examples of centralisation