This post on Best Of A Great Lot is a part of a series on the subject of designing a new form of governance. Each piece aims to stand alone, but fits together on the Table of Contents.
Note: I have put forward my own proposal for a stronger form of sortition as a useful governance tool. I call the technique SIEVE. This article does not discuss that technique.
Sortition is the practice of choosing randomly. Often the term is conflated to mean choosing political officials, but because the term means choosing randomly, there are various proposals that all go under the name. These include sortition of representatives directly, sortition of voters, and what I think is more clearly called policy juries. As far as I know nobody has seriously proposed randomly selecting policies themselves!
Sortition for political officials is an old practice, used by the Greeks during the classical era of democracy, though in that case the people who could be selected randomly were a pretty small percentage of the population. There are frequently articles exploring the idea of replacing voting with sortition in various forms, often thrown out there for the shock or humor value of the idea. Note how the author introduces the idea with "before you chuckle", which reveals the true purpose of writing this article. More seriously, Current Affairs proposes a randomly selected congress.
One of the core driving factors behind proposals for sortition seems to be how much we dislike and distrust our legislators.
As it is, our legislators are intolerable people. The fact that we elect them has failed to make them any better.
Random representation
The first form of sortition is the ancient Greek form: choosing our representatives by a roll of the dice. Here's Tim Dunlop.
I propose that we create a new chamber of parliament that I’ll call the People’s House. What makes it different it from other houses of parliament is that its members will be chosen by lottery, or sortition. Any adult citizen could at some stage of their lives be called on to serve in the People’s House in the same way that we may all be called upon to serve on a jury.
And why? Because we've lost faith in our government's ability to solve major problems:
One of the main reasons we are losing faith in our government’s ability to solve major problems is because our political system is designed to exclude ordinary people.
Here's another variant of it. The author describes some potential problems:
No pure lottocratic system has ever existed, and so it’s important to note that much could go wrong. Randomly chosen representatives could prove to be incompetent or easily bewildered. Maybe a few people would dominate the discussions. Maybe the experts brought in to inform the policymaking would all be bought off and would convince us to buy the same corporate-sponsored policy we’re currently getting.
These are some serious problems! Sadly, the author proceeds to dismiss them as comparable to the problems in our current system.
That said, it’s worth remembering the level of dysfunction that exists in the current system. We should be thinking about comparative improvement, not perfection.
Sortition for representatives has a number of problems, but before even talking about the problems, I want to highlight that it doesn’t solve some of the major problems we have in our system. Sortition won’t unbundle governance, or create justifiable governance. It doesn’t improve on the dilution of representation. It’s not at all obvious that it will drive more effective governance. The one and only thing it is guaranteed to do is create representatives who are more demographically similar to the electorate. Its advocates hope that will drive improvements, but the how is somewhat unspecified.
The most obvious objection is that good governance is actually hard, and the most important work of legislatures isn't voting on bills, it's writing them. Currently 35% of American citizens age 25 and older have a college degree. There exist tremendously smart and thoughtful people who don't have a college degree. I’m friends with several who would make excellent legislators, so I don't want to suggest that you need one to be a person I would trust with government. There are definitely reasons not to get a college degree that have nothing to do with aptitude: finances; not needing it for their chosen career; not believing it's an option, and more. But that's not the majority. Most people who don't go to college simply aren't prepared for the work expected of them in college. This seems like a perfectly understandable reason not to go to college! The work of college expects a certain set of skills and ways of thinking, primarily around reading, evaluating, and analyzing complex written material, and connecting it with everything else that you understand.
Allow me to gently suggest that the work of policy design and tradeoff analysis that's so essential to crafting good legislation isn't the sort of thing that one should do if one finds the analytical work of college to be too difficult.
It’s very possible to start trying to fix this problem by inventing eligibility requirements (you must have a college degree, etc), and I’ll talk about them in another section.
Let me paraphrase an argument I've seen frequently about sortition of representatives:
Pro: Everyday people are just as good people as legislators, maybe better!
Anti: But they don't have the skillset required to write good law
Pro: That's not important! They're good people.
Anti: How will they write good law?
Pro: Experts will advise them!
Anti: So how do we pick the experts?
A second major problem with sortition for representatives, and a problem I haven’t seen elucidated elsewhere as often, is that we’re removing an accountability mechanism and not replacing it with anything. When our current representatives go too far, they get ousted, and the person who brings to our attention just how bad a job they’re doing is the candidate trying to replace them. While you can imagine sortition systems that still allow us to vote on whether to keep the incumbent or not, the incentive to make the case to the voters for how bad the incumbent is would be significantly reduced, and many proposals don’t even describe any accountability for bad representatives.
Randomly selected voters
An alternative to randomly selecting our representatives is randomly selecting the voters.
We currently ask everyone to vote. In the 2020 US presidential vote, there were more than 159 million voters. If each voter took only 20 minutes to vote, then Americans spent a combined 6 thousand years of time to choose Joe Biden. Since Biden won by a gap of 4%, approximately 5700 of those years could be considered to be unnecessary, even more if you consider that only the vote gap in the competitive states actually matters. The chance that your vote will directly change the outcome of the election ranges from 1 in a million to 1 in a trillion, depending on what state you live in. Economists frequently argue that it makes no sense to vote.
If voters average 20 minutes to vote, it doesn’t seem like they’re evaluating candidates closely. If they're taking the time to evaluate candidates carefully, then they're spending a lot more than (aggregated) 6000 years on deciding how to spend their 1 in a million chance. It seems utterly ridiculous to suppose that citizens can make an excellent choice in 20 minutes.
Of course, voting for Congress is slightly better than that. In general, incumbents win more than 90% of the time , and by an average margin of victory of 28.8% as of the most recent election cycle. In a Congressional race, the number of votes that need to flip to change the outcome can be in the hundreds in close races, or more often, the thousands or tens of thousands, and the citizenry is only spending about 15 years per district to make a decision that 90% of the time will result in no change.
Sortition of voters could save a lot of people a lot of time, and at the same time, make voting meaningful. Select a small pool of voters by chance, say, a couple of thousand. This increases the value of them spending their time on actual research, and could reduce the number of people who vote for blind partisanship.
The actual voters would be less likely to meet the objections raised by Ilya Somin and others about the palpable "ignorance" of many voters. If, for example, they that they were part of a carefully chosen (albeit random) group of, say, 2000 Americans, to pick the next president, and if in addition there were "hearings" at which the candidates would speak and be subject to careful cross-examination concerning their views.
We could even require voters to spend time on the matter, like jurors.
In addition, the selected group can, as Sandy notes, be required (or at least encouraged) to listen to presentations on issues by political leaders, experts, and others. That could increase their knowledge still further.
It's a fair question whether this is still self-governance. One could imagine calling it stochastic self-governance or sampled self-governance, but I imagine the current complaints that our vote hardly matters would only be exacerbated by making it so strictly visible. Currently, even though your chance of affecting the outcome is less than 1 in a million, as a society, we at least pay lip service to the idea. With voter sortition, we could save so much time that the average person only has less than a 10% chance of being selected to vote in their lifetime (check my math: 2 thousand voters out of half a million, 25 elections in their lifetime, 1-(1-.004)^25 = .095).
If implemented, public pressure on the voters could be tremendous. They might well be subjected to a flood of threats, bribery and blackmail. It could be something like being the target of a twitter mob today. With high profile trials, we place jurors in protective custody and have anonymity and security procedures, but without other changes to our society, elections are likely to be much more intense than even the highest profile trials.
This form of sortition provides us some benefit against the dilution of representation, but only in the sense that the voters have more incentive to care. The problems of too many citizens to a representative still remain a challenge worth overcoming. It does little for the other problems we’ve discussed.
Citizen Assemblies
Another variation is what The Sortition Foundation calls Citizen Assemblies:
By selecting representative groups of everyday people by lottery, and bringing them together in Citizens' Assemblies ... better way to make political decisions.
Digging into their examples of current citizen assemblies, what I've found are what I would call policy juries. Someone in power decides that a specific problem, and often a specific solution or few solution ideas, needs to be evaluated, and puts out a call for interested parties, and those interested are chosen by random and brought together to discuss, learn from experts, and then make a recommendation.
Here's a recent favorable article describing their popularity.
I've been on a couple of policy juries. For a policy wonk, they're great fun, and they can be a very useful addition to our body politic. In my experience - and the notes from the examples on the Foundation's website concur - policy juries usually don't have actual power, and don't tend to be in the news much. They make recommendations, and back them up with research, and then professional civil servants and politicians take those recommendations and decide what to do with them. Between the lack of power and visibility, they attract deeply thoughtful and educated people, often current and former civil servants and politicians. I think the claims of the Sortition Foundation that they "break the hold of career politicians" and "bypass powerful vested interests" are overblown. The policy juries I've been on invited those career politicians and powerful vested interests to advise us. After all, they're the folks who know the most about the topic, and they are often very persuasive about their expertise.
Here's one fairly representative quote from one of their reports that I think sums up the current state of these:
Although the Jurors generally did not expect that their recommendations would all be implemented, they expressed a clear desire to know what had happened to their recommendations. There was a perceived lack of feedback to the Jury on the outcomes of their recommendations. This resulted in some Jurors feeling frustrated, disappointed and lacking in control and influence.
In other words, most current policy juries are very bound up within and responsive to the current power structure. Someone has to decide that a jury is needed, what to do with their recommendations, and often advise the jury how to proceed and who to talk to.
Being on a policy jury took a huge chunk of time to go in depth into a single issue. We met twice weekly for several hours for 3 months, which was the short version of the jury - some met for years - and were expected to do several hours of additional homework every week. It’s unsurprising that a bunch of policy wonks were interested in doing that, but expanding this to a lot more heavy use runs the risks of any poorly paid and semi-volunteer situation: busy folks do everything they can to get out of it, and, if mandatory, it is a much heavier burden on those without privilege. I think there’s a useful place for policy juries in a good system, but they need to be carefully thought through. Notably, policy juries as the Sortition Foundation is creating are single-issue, and so they likely suffer fewer of the problems of bundled governance.
As I was finishing this to post it, I was made aware of a new article, although I haven't reviewed it in depth. My brief skim impressed me with the sheer number of links it brings to the table! Please let me know if there are claims that I haven't adequately addressed.