Many Fools Are Better Than One Wise Person

I regularly mention, when talking about other things, that I am a great believer in “the wisdom of crowds” and democracy in the absolute.

And people are regularly baffled.

What, I see all the idiots around, and what they vote for, and how all around people say stupid things and are mean and terrible, and wrong … and I think that’s a good thing?

I do think it’s a good thing.

Unsavoury roots

In checking that my memories of the idea are more or less correct, I find it goes back to Francis Galton, who is these days more remembered as an enthusiast for eugenics. Not ideal. But he also wanted to scientifically prove that people in groups were unreliable, and was honest enough to note that the evidence pointed the other way.

People at a country fair were invited to guess the weight a cow would be when butchered. He was astonished to discover that the average of all guesses was incredibly accurate. Better than the most expert farmer or butcher. This has been replicated many times since (more usually with things like the number of marbles in a jar these days).

If you’re interested or sceptical you can find a lot about that online, so I won’t go into detail. I do find it telling that it was someone with an interest in proving the opposite who first established that large uninformed groups judge extremely well in aggregate.

So, that’s the theory

This is all fascinating, but few people are convinced that it’s a reason to trust collective decision-making, let alone voting for things that affect the real world. It seems to most people more like a weird irrelevant fact.

When it matters, they say, it’s more important to rely on people who really know. No evidence is felt to be needed for this, it’s simple common sense.

Democracy remains an important idea, but for legitimacy not decision-making. This is why we have representative democracy and why there are ever-increasing safeguards around that. It’s why “populism” is an insult. It’s why referendums are frowned on in most places, treated as a last resort at most.

As Churchill said, the best argument against democracy is a five minute conversation with the average voter. And half the population are more stupid than the average. Better to leave difficult and crucial affairs to the people who know best?

Experts aren’t very expert

I argue the opposite.

As Galton found, even with trivial judgements, a mass of inexpert people commonly get a more accurate result than any expert. And with these trivial and specific things it is possible to be an expert in the first place.

In things like politics and governance, what does it actually even mean to be an expert? Can you know enough to run a country? Can you know enough more than the average even to make a meaningful difference? Why would any individual or small group feel they know enough to be responsible at that level?

Crucially, though, people do feel better qualified than others, and we tend to go along with that.

The problem then is that people who feel they know better are also proved, in numerous replicated results, to typically be worse at forecasting than people who admit that they’re uninformed. This is not because they’re blinded by arrogance (at least, not only). When something is sufficiently complicated, detail tends to confuse rather than enlighten. So people who know little about a subject base their predictions on broadly known facts, which are actually more salient than the obscure extra facts that an expert puts weight on.

So in forecasting and prediction, experts are often both more certain and less correct. Which matters, because decision-making is based on predicting what the results of the decision will be.

… and that starts to explain a lot of politics and government.

Groups are great, as long as they’re the right kind

So the other common reaction to the apparent flaws in democracy, that we should educate voters, make them all better informed, disallow some types of people, and particularly ban some viewpoints – all these are exactly wrong.

Group decision-making works best when you preserve that magic non-expertise, and include as wide a variety as possible. If we knew in advance what decision was best then there’d be no decision to make at all, so every effort to “clean” the decision-making environment just makes it less accurate.

The key thing is that we need all sorts of wrong involved, the ones we can see are wrong as well as the ones we don’t know are wrong, because they balance. As long as we make sure all the wrongs are allowed, none of them will dominate and win out.

And as long as we preserve that sense of non-expertise. Confident groups are no better than confident people, so diversity of opinion and diversity of information is all-important. Diversity averages to accuracy.

Which tells us what we need for this to work

Firstly, group decision-making works, but mob decision-making doesn’t. So the “crowd” should be, as far as possible, a crowd of individuals.

Politics works against this, because it deliberately forms sub-groups and tries to impose common ideals, values and policies. So does “identity” organisation, where people come to feel and act as though their representation of some group is the most important thing about them. So from this perspective, a lot of modern progressive assumptions are antithetical to actual democracy, which makes sense when we see how readily those keenest on community good form groups of those considered most suited to deciding what’s good for the community. It’s axiomatic that people left to decide for themselves will decide wrong.

So, limits on group organisation.

Secondly, free flow of information is important. More information is better than less, and correctness is irrelevant.

This rubs a lot of people up the wrong way, because they see how much outright wrong information is out there in the public sphere, and how bad actors constantly push it. But the cures are worse than the disease, always, because nothing that works against incorrect information can leave the good kind immune. As we see from crowds guessing values, the average of the wrongness is usually quite reliable. Where organised attempts are made to pollute the information available, the best approach is the limits on group organisation, not what they’re organising in order to say.

So, let anybody say anything.

Thirdly, personalities spoil the process. Leaders are unhelpful.

It’s easier to vote for people than it is for particular actions, so that’s what we tend to do. But associating individual people with sets of things we’d like to happen is ineffective and misleading. By assuming that’s what democracy is, a lot of us have lost the sense that we can affect what happens at all.

So, no leaders or figureheads.

Fourthly, and most complicatedly, the decisions have to matter, and have to lead to tangible action.

People are generally extremely good at deciding in their own interests. Where this appears not to be so (I can hear the derisive snorts) we are usually unable to see their interests as they look to that person. People look as though they’re sabotaging their own futures, often, I admit, but in most cases that’s because they have little reason to believe there’s any reason to do otherwise. That aside, when people are required to decide something that matters, they take it seriously, and a collective decision made by people behaving that way is very sound.

This starts to get distorted when people stop believing their choice matters, and the results can become bizarre and counter-intuitive. Proportional representation and voting systems designed to curb extremists run this risk, because when voters find that whatever their voting pattern, the resulting governments are from the same pool of opinion, with the same assumptions about what is out of the scope of political action, those voters are inclined to turn towards whoever seems likely to smash the system just to see something, even negative, happen. At some level they don’t believe their choice will affect anything. And at another level they increasingly want to run the risk of disaster just to feel their choice does affect something.

That means that the more attempts there are at protecting the system from bad decisions, the more likely bad decisions become. Only by putting real consequence in front of the group deciding will they decide well. And decisions that might later be regretted are the price that must be paid.

So, no putting things out of democratic reach.

But surely this is all very idealistic?

Of course it’s idealistic. These are the principles I believe in, and why, not a programme I’m about to impose.

It’s why I basically trust people to make better decisions than anyone else could for them, including me, and is why when I’m in charge of anything I put far more effort into creating self-organising structures than attempt to lead. If bad decisions are being made, it’s because some of those requirements above aren’t in place.

I am also well aware that many others have radically different ideals. And I’m OK with that. As I say, we need people to be wrong in public for everything to work. But it does mean that when I have a chance to speak up in favour of more power being put directly in the hands of the people affected, I will make that argument.

Some people point to the real world as proof it can never work

And those people are wrong, too.

The most common objection is that there are plenty of examples where a democratic decision is later regretted or seems bad, and an expert might have made a different one. There are layers to answering that.

Were the requirements above fulfilled? No, almost always. That’s not a good argument, because people always say the same of the likes of communism, that failures are because it wasn’t done properly, but it does shine a light on why the decision was as it was.

But then, was it a bad decision at all? It’s important to differentiate between a bad outcome and a bad decision. If a collective decision is taken, in what sense can it be wrong? This is difficult to answer. In what sense am I able to say that my opinion is more correct than a majority?

Lastly, if decisions are to be consequential, then we must allow for some of them to be, if not wrong then at least regretted. And collective responsibility allows that to be dealt with more effectively and less damagingly.

As an example of all this, take the UK decision to leave the EU. Whether it was right or wrong in itself is debatable and will probably need the distance of history to conclude. But it was a tangible decision that affected a whole population, taken by that population. And I firmly believe that the fact it was allowed to happen, and the result followed through, makes the UK safer and more manageable than otherwise, because there is belief that people can make things happen even though it’s clear that “experts” think otherwise. If I’m right about that, we can expect more moderation in politics in the UK compared to places where voters feel more cynical.

Another objection is “the tyranny of the majority”. If we allow a majority to decide, a minority will always lose out.

Of course this must be how it works. But as long as we don’t allow the decision-making to be steered by groups, it’s acceptable. Because when the group is deciding as a collection of individuals, any majority and minority is transient and that’s protective. Everybody knows that next time a decision is made they may be in a minority, so as a majority it’s not in their interest to be heavy-handed. If there is a consistent minority then that needs attending to in other ways … but I think there is a lot of evidence to suggest that singling out a minority is counter-productive, even with the aim of protecting them, and emphasising the overall group, giving a clear stake in it, works better.

And, of course, the argument that allowing people to decide for themselves now is ruining the planet for the people of the future. This is a big one.

The problem with that argument, though, is that it’s the same idea that “effective altruism” is based on. Which goes like this: is it ethical to ignore someone in trouble because they’re further away? Should you help a hungry person in your street in preference to a desperate starving one in another country? Most people say no. But if you follow that reasoning, you can also conclude that it’s better to help people far in the future, because there are likely to be more of them and they have as much right to the help as people right now. And that leads to perverse decisions, as you can find if you read about effective altruism.

So we do clearly have a problem that current decision-making is spoiling too much, but I don’t think we can prove that allowing some wise individuals to overrule everybody would improve anything. The evidence is against that. The best I can offer is that if we made our collective decision-making more consequential and improved the total information, we’d end up with the least-bad results. People, I notice, do generally want to leave their descendants in a good situation, given the chance.

For real-world counter-examples, I offer juries and citizen assemblies. I have sat on juries, and been amazed at how ordinary people in groups can deal effectively with complicated things that matter. And countries which have put complex and divisive issues in the hands of large randomly selected groups have generally been happier with the outcomes than when their elected representatives decide. Food for thought?

A lot of this is about legitimacy

In the ultimate, I ask myself: suppose I could make perfect decisions for the people I know, and impose them against their will. Would that be a good thing?

Some people clearly think so. I respect that, but still think they’re wrong.

Life is meaningful when we have autonomy, not necessarily because everything is otherwise perfect. Even stories with no jeopardy aren’t popular. So people need the right to be wrong, even to act against their own interests.

And for me, that always applies collectively, too. A “bad” decision that affects everybody, that everybody knows they took active part in, is better than a “correct” decision they had no choice in. I’m not even sure we can call it bad – it’s likely to be better in an absolute sense, with enough perspective.

More and more I feel that I’m in the minority with this belief by now, though.