Thread by @DellAnnaLuca

THE FALSE EFFICIENCY OF CENTRALIZATION
(thread)

1/ Centralization is efficient only to the central observer

2/ This is because what's important locally is illegible to the central observer

Not because of incompetence, but bc it's impossible to be attuned to all environments

4/ For example, a fireman captain would use different clues to evaluate a "good hire" than a banker would.

This is *not* due to individual preferences (only) but to difference in environments; what's best in one environment might not be the best in another.

5/ As another example, to me and you, Mars looks dully red, as we attuned our eyes through evolution to see the variety of colors available on Earth.

To a Martian, Mars would look colorful, as he attuned his eyes to a different environment.

6/ Any skilled observer must have developed its skills by attuning his ability to spot clues in his own environment.

It's impossible to have an observer optimally attuned to multiple different environments.

Only decentralized observers can capture the richness of environments.

7/ It is impossible for the central observer to acknowledge his limits.

Just as, per the example above, you can *think* that Mars contains multiple colors, as any Martian would tell you, but you would still consider them all a shade of red and think your eyes can see more.

8/ Therefore, not only the observers must be decentralized, but they should also be given decision-making power.

It is impossible that they would truly take decisions based on the reports of the decentralized observers from the point of view of the decentralized observer.

9/ Of course, the need for decentralization does not only come from the perception problem. It also comes from a skin in the game problem.

Here, I wanted to point out how even a good-faith central observer would be unable to grasp how his decisions would affect his subjects.

10/ Therefore, the following arguments for centralization are invalid:
- If we had a honest/virtuous central decision maker it'd work
- If only we had a skilled central decision maker it'd work
- If only the central decision maker would use the right paradigm it'd work

11/ Another faulty argument for centralization is: it's more efficient.

It's not.

It is more efficient if you ask to central observers, or to people who attuned themselves to measures used by central observers.

12/ A central observer *necessarily* sees any as-planned change made by his policies as a benefit and any deviations from that plan as a cost.

To the local observer, however, costs and benefits are not related to the policy but to how the policy affected his life.

13/ As "efficiency" is largely a measure of benefits divided by costs, due to tweet #12, it is evident how central observers will always have an inflated perception of the efficiency of their central policies.

14/ Moral deviations from policies are likely to be benefits for the local observer, and legal deviations from morals are likely to be costs.

I'm not saying that all central policies are bad for the locals; I'm saying that locals should trusted on judging each one,

15/ …because, as described above, only the locals can reliably measure the effect of policies on their lives.

16/ Related thread, on the opposite of centralization

https://twitter.com/dellannaluca/status/1115526144336453632