Decentralism

Someone asked me the other day what my political beliefs are. I said “decentralism” — coined it right there on the spot, felt clever about it for maybe ten seconds, then punched it into Google. A few thousand results. Turns out the idea already has a name and a history and I’m late to my own party.

But the word fits, so I’m keeping it. What I mean by it: central control fails at scale. And this holds for all large systems — software, corporations, food supply chains, governments. As the scale goes up, the case for decentralism rises with it. Maybe exponentially.

A large system amplifies whatever’s wrong with its central node — one bug, multiplied by every connection that runs through it.

Bob down the street doesn’t like a particular race? Bob gets ostracized. His neighbours handle it. But give Bob the same character flaw and a dictator’s authority, and billions of people are affected for generations. The problem didn’t change — the scale of the system did. Scale is a multiplier, and it doesn’t care whether it’s multiplying something good or something terrible.

Canada’s in the middle of a federal election right now, and rural-versus-urban keeps coming up. Gun policy is the obvious one. The same rules just don’t work in both places — the daily reality is too different. And every time that happens, every time one rule can’t serve two fundamentally different contexts, that should be a signal. Not a signal for more argument about which rule wins. A signal that the thing is being centralized when it shouldn’t be. Ottawa keeps trying to force a single answer where none exists, and the result is strife over gun rules while more important problems go ignored.

Same pattern in food. We eat garbage shipped from central factories that maximize profit while minimizing quality — hormones, commodity corn, E. coli. Meanwhile the local farmer has a built-in accountability mechanism no regulator can match: you know where he lives. That’s not a joke. Proximity is a form of governance. The central food system replaced it with inspectors and supply chains so long that nobody’s accountable to anyone, and the food got worse.

It shows up in business and technology too. The best ideas grow from the roots — the internet is the classic case. Nobody designed it from the top down, and that’s exactly why it worked. The worst ideas trickle down from concentrated power. DRM, Enron, Madoff — one decision-maker for the whole system, and everyone else paying for their bad judgement.

This suggests a prediction I find hard to shake: long-term profitable organizations will tend to be decentralized. Command-and-control companies won’t last. The ones that push decisions to the edges, where the information actually lives, will outlast the ones that funnel everything through a boardroom. I don’t have proof of this yet. But the pattern keeps showing up.

Some people argue that centralized systems are more efficient. I think this is an illusion based on small-scale thinking. A centralized team of five works great. A centralized system of five million is a disaster waiting for a trigger. The efficiency argument scales about as well as a dictator’s benevolence.

I don’t know where the exact threshold is — where centralization tips from useful coordination into dangerous fragility. Somewhere between a small team and a nation-state, the math changes. I suspect the threshold is lower than most people think.