Our System Hands Extra Power to the People Most Certain They Are Right

Here is an internal conflict in our system of government that gets little attention. The system is supposed to weigh every citizen equally. In practice it weighs the ideologically certain more heavily than everyone else, and it does so by design rather than by accident.

The Arithmetic

The mechanism is turnout. Voting is voluntary, and the people who vote are not a random sample of the people who may.

Take the American presidential election of 2016. Evangelical Christians were, by the usual estimates, roughly a fifth of the population. Among those who actually voted, they were closer to a quarter. That is a large over-representation for a single group, and it is not a fluke of one election. The same pattern shows up wherever a group combines strong conviction with strong organisation.

The mirror image is also true. People who doubt, who are unsure whether the system represents them, who suspect their vote changes nothing, or who are simply not inclined to join anything, turn out less. The non-religious are a standing example. They are a growing share of the population and a persistently smaller share of the electorate.

So the system gives the most weight to the people with the least doubt. That is the opposite of what you would want if the goal were good decisions, and it is what the voluntary vote produces automatically.

Why Certainty Turns Out

This is not mysterious. Turnout is a function of motivation, and nothing motivates like being sure.

If you believe the election is a contest between good and evil, that your side is commanded by God to prevail, that the wrong outcome means the corruption of your children and the ruin of the nation, then of course you vote, and you drive your neighbours to the polling station too. The belief supplies both the energy and the organisation, because the same institution that gives you the certainty also gives you a building to meet in every week.

If you believe the election is a contest between two flawed bundles of policies, neither of which you fully endorse, run by people who will mostly ignore you afterwards, you might still vote, but nothing is pushing you, and the person who is pushing you is on the other side.

The doubters are, in a sense, correct about the system. I have argued at length that the vote extracts very little. But their correctness has a cost: it removes them from the count, and hands the count to those who are wrong about the system but sure.

What This Does to the Margins

The result is a permanent thumb on the scale toward the extremes.

Candidates know who votes. So they pitch to the certain, because the certain reliably show up, and the pitch has to be made in the language of certainty: good and evil, us and them, no compromise. This is rewarded, and the reward teaches the next candidate. Over cycles the whole political conversation drifts toward the poles, not because the population has moved there but because the electorate is a distorted sample of the population and the sample lives there.

It also produces a second-order effect. The doubters, seeing the conversation conducted entirely in a register they find absurd, conclude even more firmly that the system is not for them, and turn out even less. The distortion feeds itself.

The Multi-Party Objection

There is an obvious reply, and it deserves a fair hearing.

In a multi-party system the most radical voters have radical parties to vote for, and those parties mostly stay outside government. The extreme is contained: it gets a seat and a microphone and almost never a ministry. A two-party system, by contrast, forces the radicals into one of two big tents, where they become the tent's most reliable members and end up steering it.

This is right as far as it goes. Proportional systems do blunt the effect. Israel is a standing demonstration of both halves of the point: the radical parties are contained most of the time, and when arithmetic makes them indispensable to a coalition they extract far more than their share.

But containing the extreme confidence is not the same as not wanting it. The multi-party fix manages the symptom. The disease is that we have built a decision procedure in which the quality that most predicts influence is how sure you are, when the quality that most predicts being right is the opposite.

What I Actually Want

I do not want a system that penalises the religious, or any other group, for turning out. Voting is theirs to do.

I want a system in which certainty stops being an advantage. That means decisions in which the doubter has as much reason to participate as the zealot, which in turn means decisions that are specific enough to have an answer, frequent enough that a single one is not a crusade, and consequential enough that abstaining costs something real. Voting on questions rather than on tribes does all three. A question can be examined; a tribe can only be joined.

Until then, we should at least stop describing the outcome of our elections as the will of the people. It is the will of the fraction of the people who were sure enough to show up, weighted by how sure they were.