Can randomly selected citizens govern better than elected officials?

Could a random number generator like this lottery ball cage generate a better government than elections?
Alex Wong/Getty Images
Finding the best ways to do good.

If you want to guess what Congress is going to do tomorrow, just about the last thing you should do is check in with political philosophers and theorists in the academy.

But if you want to know what Congress will do in 50 years, seeing what ideas are percolating in the academy can be surprisingly informative.

That’s why I’ve been struck by the growing popularity, among academics, of a radical idea for rethinking democracy: getting rid of elections, and instead picking representatives by lottery, as with jury duty. The idea, sometimes called sortition or “lottocracy,” originates in ancient Athens, where democracy often took the form of assigning positions to citizens by drawing lots.

But lately it’s had a revival in the academy; Rutgers philosopher Alex Guerrero, Yale political theorist Hélène Landemore, and Belgian public intellectual David Van Reybrouck have been among the most vocal advocates in recent years. (If you’re a podcast fan, I recommend Landemore’s appearance on The Ezra Klein Show.) The broad sense that American democracy is in crisis has provoked an interest in bold ideas for repairing it, with lottocracy the boldest among them.

It’s a proposal that might sound ludicrous. So much discourse around “saving democracy” — including President Joe Biden’s speech calling on the Senate to change the filibuster rules — revolves around protecting voting rights and access to the polls; it feels hardly imaginable to have a functioning democracy without elections.

But there’s a reason smart people are flocking to the notion. For one thing, randomly selected “citizen’s assemblies” have shown themselves to be viable in practice (at least on a smaller scale thus far), and have already been convened in a few cases, for purposes like proposing climate policy in France or reforming the electoral system in British Columbia.

More importantly, choosing representatives randomly has some strong theoretical attractions.

The case for representation by lottery

The basic argument is that selection by lottery avoids many of the flaws and biases of elections.

In theory, representative, electoral democracy allows citizens to select authentic representatives of their interests. But in practice, this mission is undermined by the corrupting influence of campaign donors; the racial, gender, and other biases of voters; voter ignorance about which politicians and policies will best pursue their values; and on and on.

In an electoral system, a member of Congress who proposes, say, taxing imports will face a barrage of attacks from the likes of Walmart and Target that threaten their reelection.

Representatives selected by lot don’t have to campaign and don’t need campaign funds, goes the argument, limiting opportunities for corruption. As Guerrero puts it, “Lotteries excel at preventing corruption or undue influence in the selection of representatives.”

Landemore argues that lotteries can also lead to more diverse and representative legislatures than elections, which should enable better and fairer decision-making.

In earlier work, she argued that part of what makes democracy a valuable system is its ability to incorporate a wider range of information and perspectives than just those held by an autocratic elite. Lotteries, she claims, go further than elections, enabling the inclusion of perspectives systematically excluded in electoral democracy: “the introverted, inarticulate, short, and shy, as well as, typically, poor and Black or other people of color” who are disadvantaged in practice in electoral schemes.

The downsides of such a system

No idea is foolproof and one can imagine many qualms about the idea of democracy by lottery.

One is that the seeming incorruptibility of the lottery might be a function of it existing as an ideal, not a heavily contested, actually existing body like Congress.

As noted above, there are a few cases of citizens’ assemblies in recent memory, including several in Ireland (which helped advance the country’s decision to legalize abortion) and the UK (where it produced a report with ideas for reducing carbon emissions). None of these cases have involved notable instances of corruption or bribery of the randomly selected citizens, or at least no such instances have emerged publicly.

But those assemblies have typically been tasked with proposing policies that a legislature or electorate then must ratify. Ireland, as a whole, voted in a referendum on whether to legalize abortion. The citizens’ assembly’s view was not binding.

If a citizens’ assembly were given binding power to determine billions in public spending, private interests would have a massive incentive to influence the design of the lottery, what briefing materials are given to the amateur representatives, which experts get to testify before them, etc. In other words, they might be plagued by precisely the problems of representative democracy that lotteries are designed to curtail.

In her book Open Democracy, Landemore deals with this objection in depth. For one thing, in her proposal (unlike Guerrero’s), citizens’ assemblies would only propose changes that would then go to a public vote. But she also cites her experience observing the French citizens’ assembly on climate change and arguing that “ordinary citizens, once empowered, are very protective of their prerogatives and will actively and vocally resist perceived attempts at manipulating them.”

That may be, but not all manipulation is overt. Much lobbying takes the form of lobbyists providing useful information to lawmakers, albeit information framed so as to invite conclusions amenable to the lobbyist’s client. That process seems likelier to work on randomly selected citizens, who have a lower level of baseline political knowledge than people who self-select into running for Congress.

Landemore ultimately concedes that “any system would need to rely on additional accountability mechanisms, including laws regulating the role of money in politics.” That’s true enough — but it should give one pause about how likely a captured citizens’ assembly is to outperform a captured Congress, in terms of producing effective, broadly popular policies.

Second, a lot of the proposal’s effectiveness depends on being able to actually enlist the participation of a random subsample of the country. Guerrero proposes “considerable” financial incentives and providing relocation expenses and protection against firing for people chosen to serve. That would help ensure greater participation than, say, jury duty. But so long as participation is voluntary, self-selection will bias who winds up serving.

To give the lottocrats their due, such bias seems likely to be milder than any bias present in electoral democracy, which also requires would-be participants to self-select into service, along with self-selecting into fundraising, a grueling campaign schedule, etc.

People feel alienated from their government. Will this really help?

Finally, I worry a bit about lottery selection increasing citizen alienation from the political system. One of our society’s deepest problems right now is a broad decline in public trust in other citizens and in institutions — a sense that government, business, and civil society do not work for ordinary people and that activities like voting do not help.

If citizens feel like that now, imagine how they’d feel if they were given literally no choice in who represented them.

Yes, random selection means that the whole public is represented in a general, statistical sense. But this process deprives individual voters of any sense that their own actions can influence political outcomes, which could in turn worsen trust in government.

Faced with this objection, Landemore has responded that in her view, lottocratic assemblies should mostly be tasked with coming up with proposals that would then be put up for a public vote, meaning the public retains a strong say. This eases the problem, but opens up the possibility of mass corporate spending to sway those referendum results; recall Uber and Lyft’s successful $200 million campaign to pass a proposition in California to reduce their labor costs.

All that said, I’m excited to see more experiments with government-by-lottery. In California, for instance, I could imagine a citizens’ assembly empowered to produce a new state constitution that would be less encumbered by ballot referenda like Prop 13 — which sharply limits revenue from property taxes and imposes a two-thirds supermajority requirement for when the legislature wishes to raise other taxes — that make governance in that state a nightmare. I would hope such a new constitution would have a smaller role for referenda.

I could also imagine citizens’ assemblies offering a way around congressional gridlock. Suppose Congress passed a law empowering an assembly to propose changes to gerrymandering and voting rights laws — with a promise that whatever proposal emerges will become law unless majorities in both houses rejected it. Congress might not give up much power, but would enable progress to be made.

American democracy is in a bad way. Desperate times call for trying something new.

A version of this story was initially published in the Future Perfect newsletter. Sign up here to subscribe!

Back to top ↑