Against Elections by David Van Reybrouck

When friends ask me about this book, I give them a choice between 1. the spicy, controversial take and 2. the calm, well-reasoned, no-I’m-actually-serious-please-listen-for-20-minutes spiel. They invariably ask for both, and I oblige: 1. Democracies around the world are flailing, and the institution of elections is to blame. We can and must save democracy by replacing elected officials with [omitted for suspense]. 2. The remainder of this review.

The Symptoms

Against Elections begins with a critique of modern democracy. There’s some low hanging fruit in this area (see: demagogues, extremism, the other guy), but Van Reybrouck sticks to government fundamentals—legitimacy and efficiency. If the people don’t find the government legitimate, you get political tumult, general civil unrest and a possible revolution. If the government can’t effectively run the country, you get political tumult, general civil unrest and a possible revolution. The basic task of designing a government is in balancing legitimacy and efficiency. Western orthodoxy holds that this balance is best achieved through the ballot box, but we increasingly find that system under strain, squeezed from both directions.

When I first read this book in 2017, I wasn’t sold on the so-called “crisis of legitimacy”. Yeah, voter cynicism is up, turnout is down, and distrust in government is surpassed only by distrust in political parties, but a few single-digit percentage point differences in a few polls didn’t seem like the end of the world. In a complete misunderstanding of the concept of trends, I even thought to myself “Well that’s nothing new.”

Re-reading in 2023, these passages prove prescient. I started thinking about the mass of people required to make a figure move by 9 percentage points. I started thinking about what that could mean for the tail of the distribution, and I started caring a little more about the opinions of someone who has spent a lifetime studying these issues: “...an era in which interest in politics grows while faith in politics declines always has something explosive about it [...] What does it mean for the stability of a country if more and more people warily keep track of the doings of authority which they increasingly distrust?” If the aftermath of the 2020 election is anything to go by, political tumult, general civil unrest and a possible revolution.

The case against (electoral) democracy’s efficiency is easier to make. Across Europe, forming governments has become a longer, harder, and less stable process than ever before. Parties are less willing to compromise and more suspicious when they do. They prefer to grandstand—“today’s politicians can, indeed must, shout their virtues from the rooftops [...] preferably with fists clenched, legs stiff and lips together.” In a world where joining a coalition costs on average 8% (!!) of a party’s votes in the next election (up from 1.5% in 1960) none of this is incompetence—it’s natural selection.

Ineffectual governance is most striking in the US, where we routinely shutdown the government over political squabbles. Think about how stupid this is—the government (or just large segments of it) will occasionally stop (usually ~500,000 employees furloughed, a billion or so taxpayer dollars wasted, and a pause in government services) as the presiding politicians stare at each other, loudly insisting the other is to blame and demanding they cave to their reasonable demands. And for what? Forgotten political victories.

The Diagnosis

Van Reybrouck blames elections. This fits with the snappy book title and gives him edgy-iconoclast points, but I think his actual diagnosis is closer to “elections and the modern media environment interact in ways which severely misalign incentives of politicians and the broader good of the country.” This position doesn’t strike me as controversial—it’s basically mainstream. To be sure, Van Reybrouck does believe that elections themselves have significant faults, just that the present crisis is due to the interface of these faults with modern media.

The story of how commercial and social media has corrupted political discourse has been told and re-told many times, so I’ll keep this short. Commercial media (in which Van Reybrouck rightly includes social media) makes money by selling ads. More ads means more money, so there is a strong incentive to keep the audience engaged. This is best most effectively accomplished with sensationalistic drama, which the media will play up at every opportunity. Worse still, the news media is itself a political player, capable of creating actual political drama by coaxing voters into extreme positions or directly influencing politicians with (un)flattering coverage. This Vox video doubles as a wonderful explanation and example of this phenomenon (note the fast-paced, engaging storytelling, the emotional presentation, and of course the ads). All of this has the predictable effect on democracy:

Tied as it is to formats, circulation figures, shareholders and obligatory hotheadedness, the free press is far less free than it thinks and the outcome is inevitable. The collective hysteria of commercial media, social media and political parties has made election fever permanent and has serious consequences for the workings of democracy. Efficiency suffers under electoral calculus, legitimacy under the continual need to distinguish oneself, while time and again the electoral system ensures that the long term and the common interest lose out to the short term and party interests. Elections were once invented to make democracy possible, but in these circumstances they would seem to be a definite hindrance.

But it’s not just the media’s fault. Elections themselves contain perverse incentives without need for the media to goad it along. We saw signs of this as early as 1830, with Alexis de Tocqueville writing of American elections

The wheels of government cease to act, as it were, of their own accord at the approach of an election, and even for some time previous to that event [...] the head of the executive government is wholly occupied by the coming struggle; his future plans are doubtful; he can undertake nothing new, and he will only prosecute with indifference those designs which another will perhaps terminate [...] The period which immediately precedes an election and the moment of its duration must always be considered as a national crisis [...] For a long while before the appointed time is at hand, the election becomes the most important and the all-engrossing topic of discussion. The ardor of faction is redoubled; and all the artificial passions which the imagination can create in the bosom of a happy and peaceful land are agitated and brought to light. The president, on the other hand, is absorbed by the cares of self-defense. He no longer governs for the interest of the state, but for that of his re-election; he does homage to the majority, and instead of checking its passions, as his duty commands him to do, he frequently courts its worst caprices.

It could have been written yesterday! And yet, electoral democracy has had its periods of strength. Only after agitation are its worst tendencies brought to the fore.

——————

I think most people have heard some version of this story before. The value added in Van Reybrouck’s telling is not in the facts but the framing; he gives you permission to question the sanctity of elections, and in doing so frees a variable in your thinking. When the story is retold with such priming, you have another dimension to move in as you poke around the problem searching for a solution. It’s refreshing.

So what does Van Reybrouck want to do with this extra dimension?

This one weird trick will save your Democracy. (Autocrats hate him!)

Select your representatives randomly. Seriously.

This is not some part-time jury duty-esque civic chore. You quit your job. You uproot your life, move to the capital, and spend every day learning about the pressing issues of your nation, their proposed solutions, and hammering out the details with your fellow representatives. Van Reybrouck envisions a government made up of such randomly chosen citizens—a government not only of the people and for the people but also by the people.

The closest historical approximate to the proposed system is ancient Athens, circa 500 BC. There and then,[1] most government positions were filled via sortition (the drawing of lots) open to around 30% of the populace. This was how seats were filled on the Council of 500 which wrote laws, oversaw state finance and public works and broadly organized the government. These laws were then voted on by the People’s Assembly, (comprising whoever showed up to the monthly meetings) and implemented by the magistrates, most of whom were also chosen by lot (while more important positions were elected by the People’s Assembly). Jury members were similarly chosen randomly, and could moreover challenge the legality of the People’s Assembly’s decisions. Van Reybrouck emphasizes the degree and variety of citizen participation in government—

One striking thing about Athenian democracy was the rapid switching of mandates. You were a judge for one day, a councilor or magistrate—a paid position—for one year. You could not sit on the council for more than two non-consecutive terms. Anyone who felt capable of filling a government post could put themselves forward as a candidate, leading to broad participation. In fact, 50 to 70% of citizens over the age of thirty had once sat on the Council.

No less than Aristotle regards this constant switching of roles as fundamental to democracy—“The basis of a democratic state is liberty [...] One principle of liberty is for all to rule and be ruled in turn”. Here, he means to exclude elections, noting elsewhere “the appointment of magistrates by lot is democratic, and the election of them oligarchical.”

So what happened to the ancient Athenian ideal? Western intellectuals of the 18th century had no taste for it; they were avowed elitists. Jefferson believed in the “natural aristocracy among men” and that we should provide for a “pure selection of these natural aristori into the offices of government”. Edmund Burke agrees—

The occupation of a hairdresser or of a working tallow-chandler cannot be a matter of honor to any person—to say nothing of a number of other more servile employments. Such descriptions of men ought not to suffer oppression from the state; but the state suffers oppression if such as they, either individually or collectively, are permitted to rule [...] Everything ought to be open, but not indifferently, to every man. No rotation; no appointment by lot; no mode of election operating in the spirit of sortition or rotation can be generally good in a government conversant in extensive objects.

As does Boissy D’Anglas—

We should be governed by the best; the best are the best educated and the most interested in the maintenance of the laws: apart from a few exceptions, you will not find such men except among those who, possessing property, are attached to the land that contains it, to the laws that protect it, to the tranquility that preserves it [...] a country governed by property holders is in the social order; that in which the non-property holders govern is in the state of nature.

Yikes. To these thinkers, it was obvious that the everyman is unsuited to governance. Society, then, needs a method for separating the virtuous from the inept. Their solution? Elections. Jefferson agreed with Aristotle—elections are aristocratic and sortition democratic.[2]

As distasteful as the above perspectives may strike us today, it is perfectly fair to ask whether the ordinary citizen is capable of effective governance in the modern world. They are, after all, largely uneducated in history, law and public policy. They are unfamiliar with the inner workings of government and see the issues of the day through the distorted lens of mass media (if at all). Should they seriously be entrusted with power?

This is not an idle question; the entire premise of participatory democracy (as Van Reybrouck calls it) hinges on its answer.

But… incompetence?

There’s an easy way out of this: when chased by a lion, a zebra needs only outrun the slowest zebra, and likewise sortition needs only match the incompetence of our current systems to prove its worth. But Van Reybrouck has bigger dreams. He believes the people are capable of writing high quality laws truly reflective of the will of the people. Why? An experiment.

Select a few hundred people at random and pay them for a few weekends of their time. Give everyone an information packet and divide them up into committees. Each committee gets a trained moderator and access to a variety of experts. Participants then spend their time learning, discussing and debating the issues, hopefully converging on consensus.

There are only a few such experiments, but the project into which I took the deepest dive was the British Colombian (BC) citizen assembly, which was given a mandate to reform electoral law in the province, subject to approval in a referendum. I highly recommend this summary by Elena Fagotto and Archon Fung.

It is worth emphasizing that we are emphatically not searching for the ill-formed, off the cuff opinions of some rando on the street—such takes are likely to be everything our 18th century philosophers feared. The process is intentionally deliberative. Participants spend days down in the weeds of an issue with fellow citizens, reading, asking questions, and debating. The environment is collegial, and participants often express surprise at how productive and frank the discussions are. Shoni Field, on her participation in the BC citizen assembly remarks “Most of us, I remember, were thinking ‘well this is a sort of a fun exercise, but there’s no way this is going to lead to any recommendation for change’ [...] the biggest surprise was that [we] weren’t actually that polarized. We were united both in our extreme disenchantment with the current system and in a shared vision of how we would like politics to operate.”

Unfortunately such deliberative bodies don’t carry intrinsic legitimacy in the modern day, and the ensuing referendum (much closer to the off-the-street opinion) failed with 57% (<60%) of the vote. Shoni Field again—“I have now had my faith in my fellow citizens tested through way too many referendums, and they may just be too dumb for democracy [laughs] But what I saw was that when you bring together people in a thoughtful way and give them time and resources, there’s really nothing better.” Van Reybrouck agrees—

Referendums and deliberative democracy are similar in the sense that they turn directly to the ordinary citizen to ask his or her opinion, but other than that they are completely at odds with each other. In a referendum you ask everyone to vote on a subject that usually only a few know anything about, whereas in a deliberative project you ask a representative sample of people to consider a subject about which they are given all possible information. A referendum very often reveals people’s gut reactions; deliberations reveal enlightened public opinion.

This fits my intuition—separated into our own camps, we are shallow, divisive and uninformed. Face to face in the right environment, we are respectful, deliberative and thoughtful.

But ok, enough abstract theorizing—does it produce results? The main things we have to go on here are the final reports of the British Columbia citizen assembly (which recommended the adoption of single transferable vote) and a similar project in Texas which found a dramatic increase (30% +) in positive attitudes towards clean energy. It’s difficult to read these reports without recognising a basic level of competency, but the skeptic in me is unimpressed. Compared to healthcare reform or banking regulations, these issues are relatively simple. Moreover, the experiments—with their information packets and trained moderators—are set up for success, so I’m not sure how much information we gained on the raw competency side of things. Nevertheless I am impressed with how they went about the job. By all accounts, participants in both projects were respectful, listened to each other and worked towards compromise, which is all I can really ask of ordinary citizens.

I give the BC citizens assembly a B+ and recommend you mostly stick to your priors.

Citizens in power

If you believe in the basic (but perhaps not extraordinary) competence of ordinary citizens in a deliberative process, there are several advantages of placing them in positions of power over of elected officials:

(Perhaps obviously) Randomly selected citizens needn’t worry about (re-)election. Free from campaign promises and voter retribution, they can vote their conscience and be otherwise responsible. They don’t need to deal with fundraising or an arduous campaign schedule, which frees up time for actual governance.
Without a never ending political drama, (or at least exciting drama) commercial media’s incentives are less aligned with harming effective governance.
Citizen assemblies give a representative cross-section of society. This is not diversity for the sake of diversity—government is exactly the place where you want people from all walks of life voicing opinions and sharing their experiences.
The broader populace is much more centrist than our polarized parties would suggest. As in our experiments, this leads to cooperation and compromise.
Political parties would not have as tight a grip on citizen legislators. This reduces the team-sport aspect of modern politics and increases legitimacy in the eyes of the people.

With these advantages, the BC citizens assembly and the slowest zebra in mind, I am convinced that under the best circumstances, a deliberative body of lawmakers randomly chosen among the citizens of a nation would far, far outperform its current legislatures.

Critique 1: The politics of power

Public servants in the halls of power do not work under the best circumstances. There is money, there is influence, there is deception. No longer is a tweedy academic leading a discussion group in a quiet citizen assembly—now a millionaire lobbyist helpfully coaches legislators on industry regulations. The responsible representative is guarded and suspicious, bombarded as she is with constant “friends” vying for influence and not-so-subtle offers of bribes. The less responsible representatives (or the dim) are swayed by professional manipulators and enjoy a lifestyle perhaps slightly beyond their government salary. The hopeful optimism of Shoni Field gives way to a deep cynicism putting a damper on the collegial atmosphere of our experimental citizen assemblies.

For all the shortcomings of our electoral system, it places intense scrutiny on politicians. Not only in the form of neutral investigations by journalists, but also by adversarial investigations of opponents. And we still get corruption scandals! I don’t know exactly what would happen if you throw a few hundred ordinary citizens into the high pressure world of professional politics with minimal accountability. I suspect most would leave with their morals intact, but how many would succumb to the pressure? 5%? 10%? That’s a lot.

The experimental projects described earlier have friendly support structures. The real world is much less forgiving, and it’s reasonable to expect a corresponding difference in everything that makes citizen assemblies work. Van Reybrouck doesn’t address these issues at all.

Critique 2: Representative sampling

No matter how noble the cause, it is unethical to forcibly relocate a citizen and obligate them to work (non-consensenting legislators are probably ineffective anyway). Different demographics will agree to this work at different rates, so the only way to achieve a representative sample is through some type of non-random selection. All the studies and experiments did basically this—step 1: ask people randomly if they are interested, 2. Those interested attend an info session where they accept or decline participation, 3. Among those who accepted, choose the final group according to some scheme which guarantees a representative sample.

But how do you decide what counts as a representative sample? There’s a lot of room for reasonable disagreement here, as well as a lot of power. If you thought gerrymandering was tough, strap in.

Take religion for example. What counts as a religion? Who decides? How big does a religion need to be to get representation? Do you split up large religions like Christianity? If so, how? Try to imagine the amount of bickering this would cause within and between religious groups. Add an element of tribal power struggle and multiply that by every other demographic category. That’s what we’re looking at.

The world is complicated, and the exact figures you get when you ask “how many [insert group] are there?” depends a lot on how you squint at the numbers. Who exactly gets to be the Official Government Squinter is of course a matter of debate, and no matter who is picked, someone will be pissed. Whichever system is used will eventually stabilize to some kind of equilibrium (just like gerrymandering), but there is no guarantee it will be fair or equitable.

I can even imagine how Van Reybrouck’s ideas combined with intentionally poor sampling can be used as a power grab. Imagine the following hypothetical: in the coming 50 years, participatory democracy catches on, mostly among young liberals. You are a conservative political operator in a trending-left state/country. In a cunning move with broad popular support, you champion participatory democracy, and make sure to author the proposal: the legislature is to be chosen randomly; if someone declines to participate, we just pick someone else, also at random. You argue this is perfectly fair (and the public perceives it as such), but secretly you know that the elderly and religious are much more likely to answer the call of duty. Your proposal passes with broad support, resulting in a demographically skewed legislature much more amenable to your ideas. Some people notice, but it’s easy to defend their complaint to the public—“The offers are random, so every group gets exactly its share. It’s perfectly fair.”

Serious thought needs to go into the sampling procedure lest participatory democracy have a legitimacy crisis of its own. Van Reybrouck also doesn’t address this.

Thorns and viability

I don’t think either of my two complaints are insurmountable. With the slowest zebra in mind, I have quite a high tolerance for this sort of thing. What concerns me is the lack of acknowledgment (awareness?) of these issues. Participatory democracy almost certainly has more thorns than the two that I’ve outlined. I expect its leading proponents to be able to articulate them clearly, propose solutions and argue that the tradeoffs (which exist!) are worth it. This is completely absent from this book.

The Founding Fathers were intensely concerned about abuses of power. They wrote and argued and wrote and fought about how best to deal with it, and still made mistakes. Anyone seriously considering restructuring government in such a dramatic way owes that level of due diligence to the people they are trying to help. To my eye, participatory democracy doesn’t meet that threshold; it is not ready for primetime.

That said, I’m intrigued by these ideas. I do believe that informed, engaged citizens can make effective legislators and I want more people thinking about the best way to make that happen. In the meantime, I like some of Van Reybrouck’s less ambitious plans for participatory democracy working alongside elected politicians. This would likely take the form of an auxiliary legislative body to which elected politicians can delegate. As people get more comfortable with the idea of randomly selected citizens in government, these bodies can be given more and more actual power—perhaps a limited mandate to write laws on specific issues, or the power to apportion small sections of the budget. This is most effective when (as in the case of the BC citizens assembly) elected politicians want to be seen addressing an issue without actually touching it.

Of course, none of this starts with big national governments—that would be absurd at this stage—but I can absolutely imagine a few small-town politicians getting in a disagreement, and as a compromise throwing the issue to the local citizens assembly. Such institutions don’t exist (yet), but they should. They have all the benefits of participatory democracy without the high-stakes power struggles. These local assemblies would give us more information about how they play out in the wild, and would stress-test the system as they gain power.

If there’s any appetite for experimental government anywhere, this is my suggestion.

Conclusion

Barring a wave of popularity triggered by this book review, I doubt this will change the world. Politicians won’t relinquish power so easily, and if they do, there will be strings attached. But for me at least, that’s not the point. What I really got from this book was a reminder that we’re not done thinking about these issues. It’s all too easy to be cynical about government, especially when the problems are structural and all proposals to fix them are too small in scope. But that last part isn’t true! There are grand proposals to Fix Everything. They have problems, yes, but they exist! We’re still thinking about it! That conversation between Aristotle, Hobbes, Rousseau and Jefferson isn’t over, and one of these days someone is going to figure something out. I applaud Van Reybrouck for his contribution to that conversation, and heartily recommend Against Elections to anyone with an interest in government.