Sorting out sortition

Do we really need elections to count as a democracy ? No, says Belgian historian Van Reybrouck - the important thing is that we have rule by the people, not that we cast votes.

I'm going to skip over much of the preliminaries about the problems democracy is currently wreaking across the world. I have some quibbles here - I don't think representative democracy has persisted by sheer force of habit, or arose so arbitrarily. There are definitely issues better handled by experts, and it should be contrasted against direct democracy, which is impractical in a modern society. The potential advantage of representative democracy is that, ideally, it will elect people who will stand for you but will be much more informed, dedicated to the task of figuring out the solution, and so will ultimately allow you to run your life unimpeded but not disinterested in what happens in government. Choosing a representative does not automatically surrender all participation in the democratic process, and I think it's a bit extreme to depict it as such.

But let's jump in to the proposed alternative. Although largely concerned with who gets to make the decisions, Van Reybrouck is also concerned with how they make them. I like this.

A much better way to let the people speak than through a referendum is to return to the central principle of Athenian democracy: drafting by lot, or sortition as it is presently called. With sortition, you do not ask everyone to vote on an issue few people really understand, but you draft a random sample of the population and make sure they come to the grips with the subject matter in order to take a sensible decision. A cross-section of society that is informed can act more coherently than an entire society that is uninformed.

This elegantly solves the problem of partisan politics by utterly doing away with political parties. When the central hierarchy of a party is powerful enough that it can essentially dictate how its MPs vote, especially when commanding a large majority, the entire strength of group-based decision making is rendered impotent. Other proposed alternatives (such as voting for ministerial positions individually) place an enormous burden on the electorate, which sortition neatly circumvents. And of course, choosing randomly has a guarantee of getting a truly representative sample of people rather than the absurdity of only allowing the very wealthiest to become politicians.

But there's more :

The drawing of lots is not a miracle cure any more than elections ever were, but it can help correct a number of the faults in the current system. The risk of corruption is reduced, election fever abates and attention to the common good increases. Voting on the basis of gut feeling is replaced by sensible deliberation, as those who have been drafted are exposed to expert opinion, objective information and public debate. Citizens chosen by lot may not have the expertise of professional politicians, but they add something vital to the process: freedom. After all, they don’t need to be elected or re-elected.

There is much to be said for the idea that randomness, though it may not help get the very best possible choice, does helps reduce bad choices : if you can't choose wisely, choose randomly. Surely a randomised choice is far less likely to ever land upon the kind of expert psychopathic bullshitters that (all too often) dominate political leadership positions. And without a political party for them to administer, participants could focus exclusively on the decision-making process instead of internal management issues.

(I'm less sure whether avoiding pressure to be re-elected is a good thing or not. I think those who make the decisions do have to be accountable to the public in some way, otherwise that valuable freedom becomes a license to do as they please on a whim.)

Of course, randomness only reduces the lunatics, and does not guarantee that those selected will actually have any kind of understanding of the issues before them. How to address this ? He gives an example from Ireland :

Participants listened to experts and received input from other citizens (more than a thousand contributions came in on the subject of gay marriage). The decisions made by the convention did not have the force of law; the recommendations first had to be passed by the two chambers of the Irish parliament, then by the government and then in a referendum. By talking to a diverse cross-section of Irish society, politicians could get further than they could have by just talking to each other. By exchanging views with elected officials, citizens could give much more relevant input than they could have in an election or a referendum.

Of course, there's absolutely no reason that politicians can't listen to experts and citizens as well - they do. Theoretically, being professionals, they ought to do better than random members of the public... but factor in the deranged selection process (featuring a serious financial commitment, a need to be persuasive to an often uninformed audience, and an almost obscene dedication to one's political tribe) and the all-powerful party machinery, and any advantage goes out the window. I also like the mixture of procedures used, rather than just relying exclusively on one method. It's a bit surprising, then, that he doesn't elaborate on the results of that particular case. Did it actually do any good or not ?

Likewise :

Juries for criminal trials that are chosen by lot prove that people generally take their task extremely seriously. The fear of a chamber that behaves recklessly or irresponsibly is unfounded. If we agree that 12 people can decide in good faith about the freedom or imprisonment of a fellow citizen, then we can be confident that a number of them can and will serve the interests of the community in a responsible manner.

I have no idea if it's been done, but to properly evaluate this, we'd need a case-by-case comparison of trials assessed independently by multiple different methods. Otherwise, how can we say which method is better ?

What if this procedure had been applied in the UK last week? [Brexit] What if a random sample of citizens had a chance to learn from experts, listen to proposals, talk to each other and engage with politicians? What if a mixed group of elected and drafted citizens had thought the matter through? What if the rest of society could have had a chance to follow and contribute to their deliberations? What if the proposal this group would have come up with had been subjected to public scrutiny? Do we think a similarly reckless decision would have been taken?

Here we would have to factor in the enormous torrent of information coming from the media, stronger than in this example than in most others. I don't think it's at all straightforward to say what the result would be. Half of the populace is stupider than the average, after all...


Sortition has many advantages over choosing representatives by popularity contest, and I like it a lot. But it has one huge, glaring deficiency that's completely neglected in the article : it totally eliminates choice. As a replacement for elections, that's not something I can stomach. I like the feeling that I have a voice in politics, a direct influence - however small. Take that away from me and I'd be left wondering whether this new system would really be democratic at all - yes, anyone could write to the MPs, but hardly anyone would get heard because the torrent of information would have to be reduced to something manageable.

What about as a supplement rather than a replacement ? That's much more credible. Regular consultation with citizen assemblies might help, and/or they could be given some kind of limited but direct authority. More research and trials are needed to see what problems such assemblies are capable of dealing with. I highly doubt they'd be any good at economics or other specialist areas, otherwise we wouldn't need specialists. And I'm not clear from this article whether the idea is to replace the entire political body with one randomised set that deals with all topics, or if we should have multiple groups that can really focus on different areas. But that's an incompleteness rather than a flaw.

So for now I'm sticking to my idea of looser political parties with an emphasis on competitive-collaborative decision making. The selection process is important, but how decisions are reached matters too. Of course, one could always combine this with sortition, but I would rather the voters had at least some say in policy, however limited. Sortition got a lot going for it, but it needs refinement. Or perhaps I should just buy the book.

Why elections are bad for democracy | David Van Reybrouck

Brexit is a turning point in the history of western democracy. Never before has such a drastic decision been taken through so primitive a procedure - a one-round referendum based on a simple majority.