Since it appears people don't read well, (in an earlier post I said benign dictators do exist) or maybe I should say they read too well and read things that I didn't say, let me take a moment to turn to what is wrong with dictatorships.
Obviously there is a danger the dictator will be or become the wrong person, in many possible ways: sociopathy, insanity, paranoia, excess ideological fervor, lacking practical sense, or just plain brutal. Some legal mechanism for removing el Presidente is therefore recommended, but no one so far has though of a way such a mechanism could be set up without the dictator being able to set it aside.
There is another problem with autocratic government, even if it is as humane and tolerant and sweet and nice as anyone could want. When it makes a mistake it doesn't know it until things are terrible, and even then will blame others, and its efficiency permits big mistakes.
It's like an automated accounting system. Hire hundreds of accounting clerks to do the sums and addition mistakes will happen that affect one or two customers. Let a computer run it and a mistake affects millions, and does immense damage. For sure, it won't be a mistake in doing sums but in programming; the point is that power to get things done efficiently brings power to mess things up efficiently.
The accounting system analogy is well-constructed. It captures something that critics of autocracy often miss when they focus exclusively on the character of the ruler — the systemic problem of error propagation at scale.
Democratic inefficiency, which is usually deplored, actually functions as a form of distributed error-checking. The friction, the opposition, the leaks, the press that won't shut up, the factions that won't cooperate — all of that is noise from an engineering standpoint, but noise that tends to catch errors before they compound. A policy that might sail through an autocracy encounters enough resistance points that its flaws surface while they're still correctable.
There's also a feedback loop problem specific to autocracy that compounds yours. Not only do errors propagate without check, but the information flowing back to the center degrades. People learn quickly what the ruler wants to hear, and they deliver it. So the dictator's picture of reality becomes progressively more detached from actual conditions, often at exactly the moment a problem is accelerating. By the time the information breaks through — because it's too catastrophic to suppress — the error has been running uncorrected for a long time.
The Soviet agricultural disasters are a clean example. Not purely a character problem (though Lysenko had plenty of those) — structurally, there was no mechanism by which failure could be honestly reported upward without career-ending consequences for the reporter. The error-correction system was disabled by the same power that was supposed to be deploying the efficiency.
Your removal mechanism point is essentially the same problem restated: any mechanism strong enough to constrain the dictator is strong enough that the dictator will neutralize it, and any mechanism weak enough to be tolerated is too weak to function when actually needed.
I can imagine a smart dictator who sets up independent boards of experts to look at things, in fact several separate from each other competing, and other experts who pour over the reports and studies and counsel the leader. I can also imagine a leader who sets up independent data collection institutions who report data privately and separately. Without a free press, though, there will still be corruption compromising the separations that don't get detected.
Yes, and the corruption problem has a particular geometry in that setting. Each of those independent bodies, however well-designed, exists at the sufferance of the dictator. Everyone inside them knows it. That knowledge doesn't have to produce dramatic corruption — just a persistent, low-level trimming of findings toward what seems safe to report. Nobody orders it. It happens through ordinary human risk-aversion, career calculation, and the slow selection effect of who gets promoted and who doesn't.
The free press does something structurally different from expert panels. It has an adversarial incentive — reputation, sales, competitive advantage over rival outlets — to find what the panels missed or softened. Panels advise. The press publishes, and publication is the mechanism that makes suppression costly. An expert who quietly tells the dictator uncomfortable things can be quietly reassigned. A journalist who publishes them has already made the information public, so the damage to the dictator from retaliating is visible and additional.
There's also the coverage problem. Expert panels, however numerous, are still finite and directed — they look at what they're chartered to look at. A free press is undirected, which means it stumbles across things nobody thought to commission a study on. A lot of significant errors and corruptions don't announce themselves in the domain where you're already paying attention.
What your smart dictator is essentially trying to build is an information system that functions like a free press, without the political consequences of an actual free press. The trouble is that the political consequences are load-bearing. They're what make the information flow honestly in the first place.
No comments:
Post a Comment