Quadratic Voting and Policy Dollars

Link to Google Doc

By some help from professor Robert Ă–stling at SSE, I got to know about a policy suggestion closely related to my Policy Dollar suggestion. I am not sure if the idea originated with them, but the most famous proponents for this idea are Eric Posner and Glen Weyl, whom have written the book Radical Markets. (Check out Weyl’s website for more info and other ideas: http://glenweyl.com/research/)

They argue for Quadratic Voting, QV, as a way to solve the problem of majority rule in classic one-person-one-vote, 1p1v, democracy. The problem with 1p1v is that intensity of preferences does not matter, only the direction of preference matters, and thus the majority always wins. Even if there are minorities that have very intense preferences that do not align with the majority. In such a case, much well-being is lost when the minority loses. To solve this, Posner and Weyl argue that we should introduce QV, where each citizen can expend voting credits to buy votes. The cost of these votes increases quadratically, hence the name Quadratic Voting.

This system would allow citizens to express their preferences in a more dynamic way. Instead of only expressing three different preferences, yes, no, and indifferent, voters in QV can express intensity of their preferences by expending more votes either for or against the proposal.  This would allow for a maximization of the entire groups well-being instead of the well-being of one subset of the group at the expense of that of another. For this to be true though, every voter must vote proportionally to how much she cares about the issue. In the QV framework this is solved by the marginal cost of casting the next vote. The marginal cost increases proportionally with the number of votes. All else equal, if I value being able to change the outcome of a vote twice as much as you, then I will spend twice as much as you at the margin.

This relates very well to my Policy Dollar suggestion, I mentioned in my first post about it that one of the potential benefits of Policy Dollars is that it can help politicians to vote more in line with their preference. I had not thought very much about the particular dynamics of the voting, implementing QV with Policy Dollars seems sensible in that it would make politicians vote proportionally to their intensity of preference.

Posner and Weyl also propose that QV should be implemented on a nationwide level. Something which I never thought of, in part because I was solving the problem of lies and deception not the problem of preference intensity, but nevertheless it is an interesting idea. The authors estimate that the economic benefits of switching from a 1p1v-system to QV would be large, they even propose that it could be as large as 20% of national income. I am not too sure about the economic benefits, even though I think that QV would make handling public goods more efficient which in itself could bring massive value. The societal and cultural benefits could also be large, especially in the fact that the divide between politics and the public may be bridged as it would be easier for politicians to see what the public cares about and vice versa.

There have been applications of QV in settings that are more like the one I first envisioned, Riksdagen in Sweden. Where smaller numbers of people get to vote on issues by using QV. This seems to imply that QV could work in the Riksdag, although there are some issues to be resolved before that could become reality. Since income should not be a factor in voting, each politician should get an equal number of credits that can be used to buy votes. The baseline for the credits might need to be the 1p1v-system, which would mean that each politician would get enough credits to be able to cast one vote in each referendum over the year. (If there are 400 votes per year in Riksdagen, then each politician would get 400 credits to buy votes for.) If one would like to make politicians able to more clearly express their preferences, then on could adjust the volume of credits. If the credits are multiplied by four, then each politician can cast two votes per vote if she intends to vote in all votes.

As is clearly implied in the above paragraph, there are many questions that are still to be answered if Policy Dollars/QV should ever become reality. I still think these suggestions are worth thinking about, as they may bring immense societal value if implemented in the right way. It is safe to say that at least I will continue to explore these suggestions.

2 thoughts on “Quadratic Voting and Policy Dollars”

  1. My professor’s comment about one problematic aspect of policy dollars was this:

    “That once we introduce scarcity in the political process (by quantifying and limiting the amount of support one can show in favor of legislative initiatives) it might become more competitive. And I think politics is the art of persuasion and finding alliances that allow for large levels of consensus around issues. Politics (largely because of how it is portrayed by mass media) is competitive enough and lacks cooperation and dialogue.”

    This made me think, because I think he touches upon an important aspect: reaching consensus. If policy dollars increases accountability, openness, and focus on the individual politician’s values, then it might at the same time decrease the incentives for reaching consensus. The independency of the politician will hamper the need for cooperation. It might be wishful thinking, but in my ideal world I would emphasize community, and this presents a problem to that mentality and culture.

    I realize that an argument for cooperation instead of competition, is that by cooperating, politicians can lower the cost of voting – especially in this QV system. Maybe it is in fact a solution to this problem. However, the scrutiny and accesibility to how each politician votes bears effect on the public’s view of that politician, which still might be a greater hindrance to cooperation than the incentives for it. Perhaps it depends on how vocal a politician is on certain issues, which might let them “trade their votes” on other issues, and get away with it.

    Any thoughts on this aspect?

    1. Interesting! In response to the professor’s thougths: I am not sure how the policy dollar system is a worse system in terms of scarcity than the current system. The current system only allows for one vote per person, which makes that one vote very valuable and votes are scarce. The policy dollar, or QV system, would instead allow for more votes overall which would allow for more diverse voting behaviors as well as a larger payoff of persuasion. If you can persuade a single person to cast 4 votes in favor of your argument, rather than just one vote in the current system, then persuasive arguments are more valuable than ever.

      In regards to your argument on the incentives for reaching consensus. I am more favorable to your view here, because I realize that some of a politician’s constituents will not want the politician to stray too far away from his values etc. Despite that, I think that the mechanism of openness and QV can change the entire culture around politics, especially in the medial sense. Today, we view politics as a war between some set of ideas that we should not abandon, and while I think that one should not abandon one’s core ideas and values, one should still be open to bargaining to further one’s cause in politics. QV and policy dollars offers this alternative, cooperating to lower the individual cost of voting, at the cost of sometimes going against one’s values. In some sense this happens today as well. Many headlines in Sweden highlight decisions made in Riksdagen that are supported by, for example, Vänsterpartiet even though they, as a party, are supposed to be against that decision ideologically.

      Still, I think there are many cultural and behavioral aspects of a system like these that we cannot know until it has been implemented. Some of these aspects may be negative, even though I believe that an implementation would be net positive.

Leave a Comment

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.