American politics is broken. By just about any measure, we are collectively disenchanted with both the system as it exists, and the people we elect through that system to carry on its business. Particularly disheartening, money has become the defining force in shaping our political character. In Illinois, we’ll be holding a primary on Tuesday, March 18, wherein the highest-profile election will be the race for the Republican nomination for governor. The putative front-runner in that campaign is billionaire Bruce Rauner who, not coincidentally, is one of the richest men in America.
A few weeks ago, I met E. Glen Weyl at a cocktail party and heard about his quadratic voting idea for the first time. A simple idea on the surface—allow the buying of votes, but with the provision that the cost of each vote is squared as you increase in number, with all the proceeds redistributed equally to all voters—it also seems simply absurd at first. After all, money in politics is the problem, not the solution, right? But as he explained its implications—the reductive influence of wealth given the power of exponential pricing, the positive implications for the “tyranny of the majority” and so forth—I became intrigued. Why are we so blindly attached to a way of conducting democracy that was conjured up hundreds of years ago, in a time far removed from most of today’s concerns and technological capabilities? If the system is broke, why aren’t we trying to fix it? That’s what our Founding Fathers would have done.
Weyl is a rising young star in the University of Chicago economics department. Just twenty-eight years old, he achieved his first notoriety for finishing Princeton as the valedictorian of his undergraduate class while simultaneously finishing all the coursework for his PhD, which was awarded a year later. His partner in this idea is Eric Posner, a professor in the University of Chicago Law School who, among other things, is a regular contributor to Slate magazine. And yes, Posner’s father is Judge Richard Posner, the celebrated jurist and scholar.
But lest you dismiss this as just more “Chicago School” free-market propaganda, consider this: Two years ago, Posner and Weyl proposed the creation of a sort-of FDA for financial instruments, designed to reign in the proliferation of problematic securities, like the infamous credit default swaps that almost brought down the world economy. Needless to say, conservatives and Wall Street types were not enamored of this call for more government intervention.
In anticipation of this month’s election, I corresponded with Weyl and Posner about quadratic voting via email.
For most Americans, the influence of money in elections is one of the most pernicious issues of our time. Yet you’re proposing we outright sell the votes. Can you explain your quadratic voting idea and how you came up with it?
Weyl: The standard ways in which money enters politics turn the system into an auction to the highest bidder. Quadratic Voting is very different, because it makes votes increasingly expensive the more that you buy. While the first vote costs only one dollar, the second costs three, the third five and so on. (The total cost of one vote is $1. The total cost of two votes is $4.The total cost of three votes is $9.) This makes the price of the next vote you buy proportional to the number of votes you have bought thus far. Individuals will buy votes up to the point where the cost of the next vote equals the value of votes to them and as a result everyone will buy votes proportional to how much they care about the issue. Thus, rather than allowing the one individual who is willing to pay the most to dictate the outcome, Quadratic Voting adds up how much everyone is willing to pay and makes the decision in the direction of this total willingness-to-pay. It is thus a hybrid between the market and voting. Like the market, it does take into account how much people are willing to pay. But unlike the market, the highest bidder doesn’t win: the side with the most overall support does.
To give an example of how strongly the quadratic nature of costs kicks in, imagine that Mitt Romney was trying to spend his $100 million dollars of wealth to buy the presidential election. If he used it all, he would obtain 10,000 votes. This is a small fraction of a percent of the vote in the presidential election, probably not even enough to swing an election that was tied to not requiring a recount. Most estimates indicate he bought far more votes than this with the campaign advertising he engaged in. So QV is very far from allowing a rich individual to buy an election.
In fact we think it would make this harder because most issue advertising works by swaying the votes of “swing voters” or depressing the turnout of the politically apathetic. Such people don’t care much one way or the other. Under QV, their not participating would make little difference as they would be buying very few votes. Campaign advertising under QV would instead have to genuinely change the minds of people who do care, and even softening the views of an opposing individual a bit could gain a candidate a significant amount by depressing the number of votes that an individual buys against that candidate. We believe this would lead to much less poorly informed sound bites and much more substantive political discussion crossing the usually polarized sides of the political spectrum. This would be much harder to simply buy with money and would raise the level of political debate.
I came up with QV based on some work I did on eminent domain. In the summer of 2007 I lived in Rio de Janeiro with my wife, who is a Latin American political scientist. Some of the most impressive real estate in Rio is up on the hills that overlook the bay. But rather than being luxury condos, the hills are filled with shanty towns [favelas]. Why? Because Brazil does not have a good system for using eminent domain to clear such slums while fairly compensating those displaced. This brought home to me the importance of having effective policies to repurpose land.
The problem, though, is that good policies don’t really exist at present for this problem. If you try to just buy up property from each individual, there will inevitably be holdouts who will refuse to sell to get a high price. This would ruin the Brazilian slum-clearance project because, given the rampant crime in Brazil, no one would want to move into a partially cleared slum. On the other hand, eminent domain, which throws people off their land, can be completely unfair because, as famously emphasized by Jane Jacobs, it both allows the government to pay those from whom the land is taken a tiny part of what it is worth to them and, as a result, leads the government to take land that shouldn’t be taken.
A natural solution to these problems is that those selling should have a veto on the sale, to make sure it is at a fair price, but that the exercise of this veto should be determined by a vote rather than each individual being able to decide on their own to hold out. So, jointly with Scott Duke Kominers (a student of mine who is now at the Harvard Society of Fellows), I worked on developing a system like this. However, we quickly realized standard majority voting works terribly. The government would then have an incentive to give fifty-one percent of the residents just enough to get them to vote for the sale and would screw the rest. We needed a system that would not allow minorities to be exploited in this way.
As a result, we started looking at a complicated set of mechanisms invented by economists in the 1960s-1980s called “demand revealing processes.” I won’t get into exactly how they work, but they involve individuals saying how much a decision is worth to them and making payment based on what they say. These processes can be shown to always lead to the right decision if people behave according to the assumptions of the model, but they are extremely complicated and if the assumptions of the model are violated even a little bit they can lead to disastrous outcome. However, I had a student do some simulations to figure out what the payments that individuals made in these processes looked like. Every simulation we ran had the same pattern: the payments made increased quadratically in the value the individual derived from making the decision. I realized that this pattern is very general and held the seed of a system that was much simpler and more practical than the demand revealing processes.
Posner: This was Glen’s idea, but we have been working together to think about how it might be applied. In some areas, the application should be straightforward and shouldn’t offend anyone. For example, in corporate voting, it is always the case that votes are effectively bought and sold (through the purchase and sale of shares with voting rights), and QV is a variant on this. For politics, I of course realize that our proposal will be a lot more controversial. The bottom line is that the square function will limit the power of the rich to influence outcomes, and the redistributive component will compensate the poor and the middle class.
Let me give you an analogy. In Chicago, Millennium Park was partly financed by rich donors. In return for making their donations, they had (I assume) greater influence over the design of the park, and that’s why there’s a lot of sophisticated art. The rest of us had less influence over the design of the park, but we benefited because fewer of our tax dollars were spent on it. As it turns out, the park—including some of the sophisticated art—is immensely popular. QV would have a similar effect.
On the surface, it reads like a capitulation of the “99%” to the ultimate influence of the “1%.” Why would progressives potentially favor such a system?
Posner: They should keep in mind that the best is the enemy of the good. Our current system produces a great deal of inequality. QV would reduce inequality but would obviously not eliminate it.
Weyl: It’s been funny being accused of being plutocrats during this project, because Eric and I started working together on a proposal for an FDA for financial derivatives that got us branded as communists by our colleagues and was embraced by many members of the Occupy Wall Street movement. In fact, I believe that QV would achieve precisely what progressives have been trying to do for many years now.
To see why, remember that QV implements policies that the greatest number of people are willing to pay the most to see enacted. While this might seem to give more power to the rich, and would do so on some issues, whether it would depends profoundly on what type of issue we are talking about. A rich person is obviously willing to pay a lot more money to be able to lead the sort of life she wants to lead in terms of personal freedoms than is a poor person. She’s also willing to pay more for public service, like good schools and clean streets. But she’s not willing to pay more for policies that bring her a dollar of financial benefit; a dollar is a dollar and it would be foolish of a rich person to pay a lot to get back a little. On the other hand a poor person, precisely because she is poor, values money greatly.
Consider the following example. Imagine that there was one person, representing the 1%, who put forward a proposal to take one dollar from each of ninety-nine other people (representing the 99%) and give it to herself. Suppose that person spent $81 of the $99 she got this way buying votes to support it. This would mean that she would get nine votes. Now suppose that the ninety-nine other individuals, each being poor, only spent a penny each opposing the initiative. This penny would buy one-tenth of a vote for each of them, which together would add up to 9.9 votes against. Clearly such a plan could never succeed.
Thus, on the issues where progressives side with the poor, in the distribution of income, financial regulation, etc., QV would favor the 99%, I believe, much more than our current system for the reasons about campaign finance that I discuss above. But there are some issues where QV would increase the voice of the wealthy, namely issues about public services and social issues. However, these are precisely the issues where progressives agree with the wealthy. The wealthy and educated are much more likely than the poor and less educated to favor social equality for gays, rights to choice for women and restrictions on firearms. They are also much more likely to be willing to pay for improvements to the public infrastructure of cities. Their inability in the present system to express this willingness-to-pay leads many of the wealthy to move out to the suburbs in frustration with the poor quality of public services in cities. This in turn deprives the city of a tax base, exacerbates inequality and creates more social distance between rich and poor. Insulating the public sphere from the influence of the logic of the market and its efficiency doesn’t make the public sphere more equitable: it destroys the public sphere as people flee to private enclaves.
Basically, QV allows the wealthy and the poor to trade: the wealthy get the public services and social freedoms they want in exchange for the money to support food, education and housing that the poor want. This is precisely the bargain that the progressive coalition in the United States has tried to strike. And we believe it is a fair one. What is just about a society that denies the poor the right to affordable housing, to a decent education for their children and to nutritious food but allows them equal say on whether gays can marry or women can have an abortion?
Finally, on issues that are not closely related to income, progressives have always sided with the rights of minorities. Often they have had to rely on the courts to enforce such rights because standard democratic institutions enact the will of the majority. However, these court rulings often lack democratic legitimacy, undermining progressive ends. QV, on the other hand, allows minorities to express how important issues are to them in a way that will have much greater legitimacy and can be used much more easily than can the current system and therefore to win on issues that are in their crucial interests even when the majority opposes them.
We have not really discussed the redistribution of proceeds and its implications. Is that more of a byproduct for your model or do you see it as having significant effects?
Posner: It is partly a byproduct of the model but it has important implications as well. It means that people who care about a policy outcome but lose will be at least partly compensated. For example, if a town uses QV and decides to build a landfill, then people who opposed the landfill and lost the vote will receive some money. This will often be a good thing from the standpoint of equity; it will also help reduce the level of polarization in politics because the stakes of winning or losing will be reduced. The “winner-takes-all” effect will thus be reduced.
Weyl: In a large society like the United States, I would not expect the revenues generated and redistributed to be very large relative to the population size. However, in smaller groups they would play an important role in compensating people who are “moderate”…who don’t feel very strongly one way or another. In fact, in experiments that have been done surprisingly the people who benefit most from QV are precisely these moderates who don’t care much but value the compensation they receive.
How, if at all, do you see corporations and other associations involved in the process? I’m thinking of the Citizens United case and its potential implications…
Posner: Corporations would not have the right to cast votes under QV, just as they do not have the right to cast votes in our political system. They would continue to be able to make campaign contributions, and so forth—that’s just the law, according to the Supreme Court, and we can’t do anything about this. My view (Glen’s may be different) is that corporate money will always have negative effects on the political system, regardless of the voting system one uses. But those negative effects should be diminished under QV relative to the status quo because QV does a better job than the current system at aggregating people’s political preferences. An argument can also be made that corporations and other people with a lot of money would use it more productively in a QV world—by spending it to influence the views of people who are most informed and who have the strongest preferences rather than those on the margin, who are the target of ads in today’s system.
Weyl: Quadratic voting only allows individuals to participate, not groups of individuals. So corporations would play no direct role in voting. However, as long as existing law stands, they will still be able to “speak” politically. I think Eric summarizes nicely my thoughts about the benefits of QV in channeling this corporate money in a valuable direction. I would only add that there are ways that a version of QV could be used in channeling political funds that would improve the system of campaign finance. If this were done, corporations would have much less influence because they would have to pay quadratically to give money to political causes, which would reduce the amount they are effectively able to give.
Is it better suited for subject-based votes (referenda) than candidate voting?
Weyl: QV could work equally well in both systems and might even be a larger improvement in candidate voting. In candidate voting, when there are more than two candidates, things can go famously wrong: Ralph Nader can tip the election to George Bush, for example. In fact, there is a famous result in economics called “Arrow’s Impossibility Theorem” that shows that there is no fair way to run a multi-candidate election using traditional voting. QV can resolve this issue and give a fair way to decide among many candidates.
Posner: I think Glen and I disagree here. I can see it working well for referenda, without requiring much change in how they are conducted. If QV were applied to candidate voting, it would require radical changes in law, including probably constitutional law, and how people think about elections. This is a long-term project.
I can’t think of a more appropriate hypothetical application of your idea than a Republican primary, such as the March 18 gubernatorial primary in Illinois, where the wealth of one candidate, Bruce Rauner, has been a constant topic leading up to the event. Hypothetically speaking, if you ran the election and used your system, how might it work in this case?
Posner: I haven’t been following this closely, but the bottom line is that QV won’t give much of an advantage to a rich person because votes become very expensive thanks to the squaring function, and the population is so large.
Weyl: Honestly, I don’t know enough about the race to give any details, but the advantage of QV in this case is that individuals can buy votes both for and against each candidate. This avoids the problem of “spoilers” and vote splitting. These problems have led to disastrous outcomes in the history of democracy. Hitler won in 1932 largely because a number of middle-class Germans were so afraid of communist rule. If Jews had been able to vote against Hitler and the German bourgeoisie against the communists, likely a centrist government would have been selected. I doubt any such extreme issue would ever arise in Illinois politics. But one can imagine the anti-Tea Party vote being split while the Tea Party vote being united and this leading to the selection of a radical candidate. QV would avoid such perverse outcomes as those who are scared about the Tea Party candidate’s electoral prospects could buy negative votes.
Though you two work for the same university, you’re from different departments and different generations. How did you come to collaborate?
Posner: We got to know each other and discovered numerous overlapping interests as well as a relatively similar world view—which I would describe as pragmatic and empirical rather than ideological, and anxious to explore new ideas rather than to revisit old debates. We also want to improve public policy, and are willing to go through the time and effort to talk to government officials, write for the media, and so on, in order to advance this goal.
Weyl: Just to be a bit more concrete, Eric and I met because I like talking to folks in other departments so emailed him to get lunch. We found immediate intellectual chemistry on a number of levels and have grown closer gradually since then.
Discuss briefly your plans to also turn this into a business.
Posner: We have formed a business called Collective Decision Engines and are exploring various ways to commercialize QV. For example, QV can be used by businesses that rely on employees to make collective decisions (for example, hiring, scheduling meetings, and deciding which projects to pursue) or that consists of stakeholders who collectively determine policy (for example, partners in a partnership, members of trade associations, even shareholders of corporations); for market research; even for Internet-based gaming and other activities where groups of people make collective decisions. We are developing software that will enable groups to us QV easily and efficiently.
Finally, this proposal has progressive implications as does your “FDA for finance” idea. Is there a new “Chicago School” emerging distinct from the famous more market-doctrinaire legacy of Milton Friedman, George Stigler and, to some extent, your father, Eric?
Posner: What we share with the old school is an interest in market mechanisms. QV is a way of “marketizing” policy outcomes. Yet I (and I think Glen) are less skeptical of the government than the old Chicago School was, which makes us more open to regulation. Moreover, our normative stance is utilitarian, which leads us to be more sympathetic to wealth redistribution. The old Chicago School people might better be regarded as libertarians. Ironically, utilitarianism is truer to the premises of economics than libertarianism is.
Weyl: I agree with Eric on all of this. Speaking more broadly about Chicago, I think that both Eric and I are closely attached to what we really believe is the core of Chicago: an atmosphere of intellectual engagement, of taking abstract intellectual ideas seriously enough to ensure they really work in practice and the “price theory” approach to analyzing market that considers how large numbers of individuals, rather than a few people playing a game, interact to create social order. No one at Chicago has ever asked me about or judged me based on my political views and I think throughout the university economists and those interested in economics from other fields have a diverse range of views, despite our reputation.