Are You Too Stupid to Vote?

The political philosopher Jason Brennan did an AMA here 9 years ago. He argued, among other things, that a shift in the direction of epistocracy could create a much better political system than we have now. Over the past decade, I think his case probably got considerably stronger (from a few obvious angles) in the minds of the highly informed nerds who dwell here. But is this case, all things considered, good enough?

Excerpts from the article:

In 2016, the Annenberg Public Policy Center found that only 26 percent of Americans could name all three branches of government. Thirty-one percent could not name a single one. These are the people choosing the leader of the free world.

If that sentence pissed you off, congratulations: you have just felt the emotional pull of epistocracy, and you should be suspicious of it.

Epistocracy is the idea that political power should be distributed according to competence. The smartest version of the argument goes like this: political decisions are high-stakes. They determine who goes to war, who goes to prison, who gets healthcare, who starves. Decisions of this magnitude should be made by people who have some idea what they're doing. Democracy, by giving every adult an equal vote regardless of how spectacularly uninformed they are, systematically violates this principle. So maybe we should try a different arrangement.

The philosophical case for epistocracy is almost embarrassingly old. In Book VI of the Republic, Plato describes the "Ship of State," an allegory so effective that people are still deploying it 2,400 years later. The setup: imagine a ship whose captain is large and strong but slightly deaf, a bit nearsighted, and ignorant of navigation. The crew members brawl over who gets to steer, flattering the captain or drugging him to seize the helm. None of them know how to navigate either. The one person on board who actually understands the stars and the currents, the true navigator, is dismissed by the crew as a useless "stargazer."

Plato's point is obvious: governance is a techne, a craft, like medicine or shipbuilding. You would not let passengers vote on your surgical procedure. You would not poll the crew on whether to sail into a hurricane. Why would you let the ignorant masses steer the state?

Brennan sorts citizens into three categories. Hobbits are the politically disengaged: they know almost nothing about politics and care less. Hooligans are the political junkies, but their engagement is tribal rather than truth-seeking. They treat politics like a sport, cheering for their team and processing information through a thick filter of motivated reasoning. And then there are the Vulcans: rational, dispassionate, well-informed citizens who evaluate evidence without tribal loyalty.

From this empirical base, Brennan derives his competence principle: "Citizens have a right that any political power held over them should be exercised by competent people in a competent way." In practice, this means that high-stakes political decisions made incompetently or in bad faith are presumptively unjust. He draws an explicit analogy to jury trials. If a jury convicted someone out of ignorance, malice, or whimsy, we would not accept the verdict as legitimate. We believe the defendant has a right to a competent tribunal. If that's true for criminal trials affecting one person, why shouldn't it be true for elections affecting hundreds of millions?

But every advocate of epistocracy thinks the wrong people will be excluded. Libertarians imagine an epistocracy that produces libertarian policies. Progressives imagine one that produces progressive policies. Technocrats imagine one run by technocrats. Nobody imagines an epistocracy from which they would be excluded. This should be alarming. If your proposed system of government conveniently coincides with your own group interests, that's self-dealing.

The epistocrat's fantasy is always the same: a competence test that conveniently selects for the kind of competence the epistocrat already has. A political science professor imagines a test on political science. An economist imagines a test on economics. A lawyer imagines a bar exam for voters. But the nurse who understands what happens when Medicaid gets cut, the farmer who understands what trade policy does to commodity prices, the former prisoner who understands the criminal justice system from the inside, these people have knowledge that no exam will capture and no epistocrat will think to test.

The Condorcet Jury Theorem, originally formulated in 1785, proves that if voters are on average more likely to be right than wrong about a binary choice, and if they vote independently, then as you add more voters the probability of the majority being correct approaches certainty. This is a spectacular result if the premises hold. It means that a large enough democracy is essentially infallible.

The catch is the word "if." The theorem has a dark twin: if voters are on average more likely to be wrong than right, then adding more voters makes the group more reliably wrong. Bryan Caplan, in The Myth of the Rational Voter (2007), argued that this is exactly the situation we're in. He identified several systematic biases in public opinion: an anti-market bias (people underestimate the benefits of markets), an anti-foreign bias (people overestimate the costs of trade and immigration), a make-work bias (people overvalue job creation relative to productivity), and a pessimistic bias (people think the economy is doing worse than it is). If Caplan is right, the Condorcet theorem says democracy will systematically amplify these errors. The bigger the electorate, the worse the outcomes.

I think Caplan overstates his case a bit (an irony he would appreciate, given his arguments about overconfidence). His catalog of biases assumes that economists are right and the public is wrong about the effects of trade, immigration, and markets. That may be true on average, but it smuggles in a contestable premise about what counts as the "correct" answer to policy questions. Is the correct immigration policy the one that maximizes GDP, or the one that preserves social cohesion, or the one that respects human rights? Economists might have useful things to say about the first question. They have no special authority on the second and third. And the whole point of democracy is that these tradeoffs are not technical questions with right answers. They are value questions that require input from the people who will live with the consequences.

Democracy is epistemically mediocre. The median voter is badly informed. The electorate is systematically susceptible to demagoguery, tribalism, and motivated reasoning. The policies that emerge from democratic processes are frequently incoherent, contradictory, and worse than what a panel of experts would produce on any given issue. If you assembled a commission of the best economists, epidemiologists, and trade analysts, they would probably produce better monetary policy, better pandemic response, and better trade agreements than the United States Congress has produced in the last fifty years. This is Brennan's point, and it's largely right.

The case for democracy has never been that the people are wise. The case is that nobody is wise enough to be trusted without accountability, and accountability requires power to be distributed broadly enough that the rulers cannot simply ignore the ruled. A lower bar than "the people always choose well," yes, but a more honest one: whatever the people choose, they can unchoose it, and the people who chose it have to live with the consequences. That constraint, weak and messy as it is, produces better long-run outcomes than any system that removes it.

Plato's navigator knew the stars. But Plato lived in a city-state where women, slaves, and metics had no political standing, where only a fraction of residents counted as citizens, and he still thought democracy was too dangerous. The epistocrat's error is always the same: mistaking a question about values for a question about facts, and then concluding that the people who know the most facts should get the most power. Politics is not celestial navigation. The stars are not fixed. The destination is contested. And the passengers have a right to help choose where the ship is going, even if, especially if, they cannot plot the course themselves.