Quadratic voting (2014)
ericposner.com
ericposner.com
If I understand correctly, some emergent effects would occur.
1. The poor would collectively have less voting power.
2. The poor would likely become less poor due to distribution of voting funds. This would be roughly proportional to the amount of voting power lost.
3. The middle class, assuming it still exists the way we know it, may or may not have increased collective voting power.
4. Upper-middle class will gain proportionally more collective voting power than what middle class gained, but they are fewer in population. The 1% will gain even more, but they are only 1% of the population. Each social class (aside from the poor) will have more representation, while the poor will receive money (and a tiny amount of representation).
5. More passionate voters will pay more, and polarizing or strongly motivating issues will create a less-predictable situation at the polls. This could be a very good thing or a very bad thing, I don't know. Perhaps this is where the magic happens.
6. Political groups will form voting coalitons to vote certain ways, and may be paid to do so, beyond the voting fees, unless legally prohibited. However, this may prove to be not the most viable use of resources.
I think that this could work in positive ways. However, I wouldn't bet on it, as there are as many things that could go wrong. When dealing with complex strategy or architecture choices with unknown consequences, it is a hard sell to willingly go with one that brings significant risk. However, one can argue, especially with the current political climate and lack of faith in the system, that not switching to this would be the bigger risk.
I'm interested in finding smaller scale applications of such a system, as a crude proof of concept.
Disclaimer: I'm not a US citizen.
It's actually more cost effective to buy people's votes the usual way.
If 1000 votes cost £1 million, and 100 wealthy individuals all decide to spend that much in an election, that's 100,000 votes when these individuals would only usually have 100 votes. That's going to have an impact, especially on local elections.
As for it being more cost effective to buy votes the usual way, the "usual" way relies on voter fraud and massive advertising campaigns, one of which is illegal, and the other is very expensive and the voting decisions are not guaranteed. On the other hand, quadratic voting would be legal and guaranteed.
One of the things you're missing is that people who can afford to buy that many votes for a single issue are very rare and could easily be out voted by a small organized group that buys multiple votes, but with a lower power multiple.
Assuming all votes cost $1, then 1,000 votes costs 1mil. That could easily be nullified by 100 voters buying 10 votes.
Consider also that the money would be rebated equally. So you might spend $100 and get back $250 and win the vote.
Do you not see the problem with that? With the situation you describe, it takes the coordination of 100 individuals to nullify the actions of 1 wealthy voter. Do you not see how much easier it is for the wealthy voter?
I believe this is what is meant by "the collective decision rapidly approximates efficiency as the number of voters increases".
How balanced this is, I'm not sure. Why use the square of the number of votes rather than the cube? But it isn't immediately clear that this gives a few rich people significantly more voting power in a situation where a larger number of poor people feel strongly about an issue.
Two points:
1. You've chosen a value that's too small for the wealthy voter. $1000 is the cost of eating out at an expensive restaurant (e.g. https://financesonline.com/10-most-expensive-meals-from-mich... ), not the amount someone with money will spend to gain influence in an election.
2. There are cheaper ways to multiply the vote. Let's assume quadratic voting was in place in the 2016 US Presidential Election, and you're a wealthy person who wants to see the Democrats win. You have $1,000,000 you're willing to spend. What's the cheapest way to maximise your influence? Reach out to other Democratic voters to give them money to multiply their votes. On your own, the $1,000,000 may only net 1000 votes, but spreading the $1,000,000 around to people who were going to vote Democrat anyway could see that number skyrocket. Let's say you give $100 to 10,000 voters, now you're looking at 100,000 votes.
Even if those without the most money win, quadratic voting stacks the odds against them.
Yes to make it simple to intuitively see how rich people paying more gives more voting power to everyone else IF they choose to spend it that way. If you want to think about US presidential elections, you'd need to increase the total number of people by 3,000,000. Based on my math, this would just further amplify the effect I was trying to illustrate with my numbers.
> Even if those without the most money win, quadratic voting stacks the odds against them.
As I understand the math, it doesn't necessarily do this. The goal was a voting system that takes into consideration not just how many people want a particular outcome, but how badly they want a particular outcome. Today the way that is done is by donating money to be spent on advertising to sway votes or get people to go to the polls. Quadradic voting is designed to be more fair than the current way of allowing people to magnify the strength of their preferences for a particular outcome.
My initial reaction was the same as yours. After looking at how the math actually plays out and considered how money is currently spent on influencing votes I think it is probably a lot fairer than it looks at first for the goals it is trying to achieve. (Still could be a really bad idea for other reasons.)
I respect that you had good intentions, but I think the approach was flawed.
In your example, $1000 vs. $400 ($4 * 100), what does it take for the rich person to gain the lead again? They can either spend more or split the money. For example, if they split the $1000 with a family member who was voting for the same party/candidate, that's all it would take to gain the lead.
Also, consider the difference in fundraising potential between the rich and the poor. For example, in the 2016 US Presidential Election, Bernie Sanders ran what I believe to be the most successful citizen-led fundraising campaign in US political history, raising approximately $228 million:
https://www.opensecrets.org/pres16/candidate?id=N00000528
In contrast, in the same election, just the top 23 super PAC donors alone spent a total of $468 million:
https://www.washingtonpost.com/graphics/politics/superpac-do...
So the most successful crowdfunded campaign in history was outspent by just 23 people.
I don't know if you were a Bernie supporter or not, but I hope you'd agree that he was a candidate that people were strongly behind. However, due to the income inequality in the US, the difference in spending power is massive. It's not a matter of the general public wanting something badly enough, $228 million is a clear message that the want was there.
The only caveat with all this is that not all the super PAC money would've been spent for a single party. However, the majority would've been spent supporting corporate-friendly candidates.
http://www.huffingtonpost.com/entry/super-pac-donors-2016_us...
"An analysis of 2016 super PAC filings shows that the 257 donors giving $500,000 or more have contributed a total $709 million."
Somewhat true for elections of political representatives, but even more true for things like Senate Committees - these require party membership fees that get paid, which come via donations from lobbyists.
Not party political, both sides do it: http://www.definingthemachine.com/party_dues.html
Unless one can come up with a way to inflict the same budgetary pain on each individual for each vote-expenditure, I don't really see how you would argue for its implementation/superiority in practice (against the other forms of voting out there I mean, not necessarily contrasted just against first past the post...).
Though this all depends on what one means by "optimal voting" or "efficient", since I don't believe there can be an objective definition of such.
I could be wrong, but I prefer the possibilities of such a system not directly involving money. The distribution of wealth is so imbalanced that I cannot see buying votes to have any kind of positive impact.
The author of the papern should go visit North Korea or Saudin Arabia and perhaps live there if he thinks the majority is tyranising him with their voting rights.
So, fine, while the current form of democracy may have somewhat worked until now, that definitely does not imply that we reached an optimal voting system. If we can make the world a little bit better, why not?
(This paper looks interesting, but, of course the devil is in the details of implementation; would I personally want this system? No idea.)
People were governed and had different levels of say through the history.
Perhaps the most extreme example would be the French Revolution where the aristocracy had more votes per person than the common peasant, and we know whose heads rolled afterwards.
The problem isn't game theory. The problem is that power attracts certain personalities that are not altruistic. The current iteration of having say in common decisions is explicitly trying to curb the maniacs from reaching/staying in power.
> If we can make the world a little bit better, why not?
Sure. But right now, will Trump sing Kumbaya, or is USA regretting the decision? If it was 40-40, imagine him buying 20% of the votes.
"Many forms of Government have been tried, and will be tried in this world of sin and woe. No one pretends that democracy is perfect or all-wise. Indeed it has been said that democracy is the worst form of Government except for all those other forms that have been tried from time to time.…" - Churchill
CGP Grey has made an excellent and well known video series explaining this, so I don't think I need to debate it much further here.
CGP Grey Voting Series: http://www.youtube.com/playlist?list=PLej2SlXPEd37YwwEY7mm0W...
OK, here you go: "Weyl and Lalley prove that the collective decision rapidly approximates efficiency as the number of voters increases. By contrast, no extant voting procedure is efficient."
Completely on point. Especially true of the American public.
Getting people to migrate to this scheme is a huge problem. Today in America, voting is a right. However, to many, it appears to be a financially free activity as well. Or some may tie it in with "you get to vote if you pay taxes". Regardless, this "pay into a fund in order to vote" won't really go well -- we already pay a fund (taxes), and it's distribution is mishandled / non-optimally allocated / gamed by people in the system.
I realise of course, there is a world of difference between pointing out there is a country with a different/more complex voting system, and migrating a country with culture/history of a given voting system to a more complex alternative...
1. Instant runoff voting (preferencing the candidates from 1 to x).
2. Compulsory voting (actually a fine for not visiting the polling station and getting your name crossed off, voting is optional).
The two systems together work to keep everyone involved to a certain extent in the political system, a choice in who you can actually vote for, and historically, stable centralist governments. We have over the last decade or so ended up with less stable governments, but this is more a reflection of a convergence to the centre. The differences between the major parties is now so small that changing does not make much difference in policy.
The one downside is it does make Australia a rather boring place to live :)
Basically, you cannot have radical societal or cultural changes because you have to get everyone on board ahead of time.
In a two party system and optional voting, you can have only a small fraction of the population who particularly want some changes actually get those changes. Also, this strategy can be used to destabilize a system away from local maximum (if we are searching for some better state).
A really good current example is same sex marriage. Australia is one of the last western countries to allow this (should be implemented before the end of the year), but it now has the support of 70% of the population. I would rather see change occur that lasts, than a small minority impose their will on the majority and having to deal with the resulting backlash.
Granted a lot of this is about stickiness of laws but it seems like a huge battle has to be fought for removing systemic inequalities, when that seems to be a major problem in the current system
It's only very recently that reproductive rights have broad majority support. The Civil Rights Act could never be enacted if it had required 70% support.
Once the battle is won then these rights get protected. At a time when systemic inequalities are at the forefront, inhibiting change seems like something that would go against the general political currents.
It is worth debating if the civil rights act should have been enacted before it had popular support. Rather than undermining the basis of suppression, imposing something on the majority without their support brings a resulting backlash that can last generations. We only have to look what drove the last 50 years of politics in the US south to see the result. From my comfortable perspective it might have been better to let the south come to its own opinion (with a lot of outside nudging) that things really had to change.
Its hard because laws that ensure basic civil rights go through the same process as tax law, but one is vitally more important for a better future for all
Yes relying on achieving popular support before acting will always mean that change happens later than it really should, but I think this price is worth paying to ensure long term support.
I do think political systems that reguire building broad popular support are desirable in general because they encourage openness.
You basically cannot get out of it unless you can convince most people that it's a good idea. With a 2-party system you can get someone who represents those people to rule for a while and maybe correct the course.
Also, with optional voting, we basically delegate decisions to those who really care about something. Having lived in a country with mandatory voting, I saw a few elections where people's views and the results of some elections were easily manipulated by false news close to the election day. People who are ambivalent I think are more likely to change their decision over such things.
I don't know about boring, from my perspective it's more comical. Who's turn is it to be Prime Minister this week? and can she stack it in high heels on the grass again. We've had the same PM for nearly a year, must be time for another spill.
Your post did make me LOL remembering Julia trying to walk on grass in high heels and then faceplanting.
That's not a bug, it's a feature. There's a reason that "may you live in interesting times" is considered a curse.
It's the change that make it hard to implement and it's not just in politics but in most things humans are involved with.
We accept what is as the norm because we grew up with it.
Let's say I'm a very wealthy person, and I have $4M to spend. It looks like I can buy 2000 votes. However, money is very fluid. So I decide I'm going to find 10,000 "budget-constrained" voters and give them each $400. I'll say, "Here's $100 for buying 10 votes, and keep the remaining $300 for yourself." Now I've just purchased 100,000 votes.
Ah! ZenoArrow (https://news.ycombinator.com/item?id=15206589) beat me to it. Well done.
Which is what prevents this today.
If real money is allowed, then the very poor are completely disenfranchised. Many don't have enough money to put food on the table, never mind spending it to vote. That gives the wealthy a huge advantage - more so than they have today.
It sounds like you are saying that a system that allows you to buy other people's votes skews toward rich people. We don't outright allow that today unless you count advertising. With the quadratic system proposed, rich would have to choose between advertising or putting more money into the voter pool (is distributed to everyone and thus increases the voting power of the poorer majority).
Here, you just buy extra votes for a block of people you have good reason to believe already intend to vote your way, and the net result is very much in your favour.
The only way this works is if you're giving everyone credits to spread between multiple known issues according to their priorities (and even then interested parties can theoretically trade votes on the side)
Calling the majority tyrannous and assuming fringe interests to be naturally benevolent reeks of an agenda.
Please don’t insinuate bad intent where there needn’t be any.
An enlightened millionaire minority, by the looks of it.
It more often refers to settings where a majority and a minority have conflicting interests, and in which the majority is inclined exploit the minority in ways where the damage endured by the exploited far outweighs the benefit to the beneficiaries. The phrase has been uttered: "Democracy can be like two wolves and a sheep deciding what to have for dinner."
EG:
- A majority of city-dwellers will vote to destroy the homes of the forest people to improve the development of their cities, however slightly.
- A majority of old people will vote to tax the young more heavily to achieve a transfer of wealth, even though the gains of each beneficiary is smaller than the loss of each loser
- That same majority of old people, finding that their preferences in nearly every domain are different from those of the young, might vote to raise the voting age.
- ...Look beyond the surface level. If you can't trust the majority to make the right decisions, who do you trust to do so? The only reason to bring up the tyranny of the majority is to imply that another (smaller) group could do it better.
I'm not saying it is a good idea overall. Just that while it looks unfair when just consider how the votes are purchased, things look much fairer when you consider that the money used to purchase votes is distributed evenly back into the system.
I'm not sure it would work well for the US government, but it seems like it might work very well for other organizations--for example stockholders in corporations.
Now: Your vote can be bought once Then: Your vote can be bought arbitrarily often with diminishing returns.
In many European countries ethnical minorities are given a few guaranteed seats in the parliament to ensure that their voices are heard. The same could be done for a wider spectrum of minority groups that are of interest to society, selectively boosting their political influence, but still filtering out the crazy ones. Problem with this is that it's not as stable as giving everyone the same rights unconditionally, as it's much more prone to manipulations, like cutting deals with minority representatives who got elected without voting and making wide coalitions with them in order to take over the power, against the will of the majority.
As long as one group's actions are considerably detrimental to another, that other group will care.
In a way, A can blackmail B, because the money gets distributed later (A will get a lot more back than they put in, because B bought an unnecessarily high amount of votes).
Is this scenario acceptable?
Also, there should be room for both rationality and emotion in public debate. Emotion is a key part of the human experience, making decisions based on rationality alone will stop us seeing why certain rational decisions should not be followed.
> Hmm, but what if "a minority that cares passionately" are nazis or religious fanatics or some other extremist group?
Then they cause an outcome by paying all the people that are not in the minority lots of money, and then next election the majority uses that money, more effectively per voter, to reverse the outcome.
If you think the median individual is rational enough to be allowed to vote at all (I don't), then the passionate minority still has to keep the majority happy or else they transfer large sums of money and therefore a larger sum of distributed-voting-power onto that group for future elections.
Understanding the quadratic effects and that the money is returned per capita is absolutely key to the beauty of it. Imagine 10 rich voters buying an election out of 1000 voters. They spend $10 million to do so. The next election, those 990 (now unhappy) voters, with the proceeds of the $10m money they received, can all buy many more distributed votes than the 10 voters were able to buy, and reverse the decision overwhelmingly.
So if you're really worried about this (and think voters are rational), just allow snap elections or votes of no confidence if there's enough protest, and the majority with their new dollars can right the ship very quickly.
It's all bread and circuses, with the media running the circus. Just spend 10 minutes watching one of the big networks and you quickly realize it's all just rhetoric and bias, sprinkled with facts. No one reasons or argues, it's all just emotions and word-juggling. E.g. the recent "undocumented citizens" and "dreamers" double-speak when referring to illegal alien children under DACA. They're children, they're illegal aliens, it's not that hard.
Edit. Grammar.
I've long found it an interesting idea to allocate limited capital assets (aka land) in a similar manner, select best use (aka owner) through highest bid (aka vote), but offset the opportunity cost, and rent extraction, by expressing said bid in a rent paid as a public dividend.
Perhaps when discussin public policy decision by means of voting with your wallet it opens the door to also discuss market decisions as a democracy implementation.
(Oh well, I have a 58-page paper to read to see if this comment was entirely of topic)
Fix either one, preferably fix both, and I suspect people will stop proposing brain dead ideas like this.
I'm assuming they're working within some mathematical formalism where efficiency is defined in some manner, but I can't access the linked paper. Can anyone explain what efficiency means in this context or provide some links to the mathematical framework for this stuff?
I love that there are so many better ways to vote, and I hate that we still use FPTP.
I'd assume that this paper is written by someone who has intimate knowledge of AIT, just looking for some perspective.
- Votes with at least three options. A simple majority vote between two options does in fact satisfy all of Arrow's fairness criteria (if only vacuously).
- Ranked voting systems, i.e. ones where voters only give a rank ordering of their options, not any more information.
Quadratic voting is neither of those: It applies only to binary decisions, but for those decisions, it tries to be more fair than Arrow's criteria call for, by getting out not only information about which option voters prefer, but how strongly they prefer it. Whether it could be extended to more complex votes than binary ones, I don't know.
What does the general population think if aristocracy is about to buy back it's way to power?
And it's of course a cheap take on the tragedy of the commons.
If there is say - Aerith and Bob running for president, say there are Charlie and Drako that want to buy two votes each for Aerith.
Do they each pay 4 or do they pay 16?
Otherwise, a marginal vote purchased by a different voter would change the total price of prior purchased votes. If votes were tallied per candidate, voters would bid without knowing how much they owe until after the election.
But as the paper notes, it is mainly around referendum voting. And referendum voting can be viewed as opposed to representative democracy. Voters will happily vote yes on one referendum for higher benefits, and yes on another referendum on lower taxes, and the end result is laws that don't make sense. This can be seen as an example of the Condorcet Paradox [1]. But the point is, referendums are generally a terrible way to make decisions.
So on page 45 of the paper [2], the authors discuss that democracies are generally run according to representation, not referendums. So first they suggest using QV to elect candidates. Yet when given a choice between 2 (or even a handful) of candidates, do we really expect QV to make a difference? Republicans and Democrats are both extremely passionate, so I'd call it a wash. There's such a mix of social and economic issues large and small, QV doesn't seem to help untangle that.
The second thing the authors suggest is for representatives to "put their constituents' money at stake" when voting on laws. The authors wisely don't go too deeply into this, because it's clearly nonsensical when representatives are from a district which contains significant members of both sides. They actually admit as much ("QV in representative assemblies would probably make more sense in a parliamentary system than in a presidential system") but actually get their terminology wrong -- it would make more sense in a proportional representation system than in a first-past-the-post system. (Presidential/parliamentary refers to the relationship between the legislature and the executive, which this proposal has nothing to do with -- which reveals their amatueur-ish level of analysis in this case.) At least they're honest in saying "many details clearly need to be worked out".
But I have an even more basic question, even for deciding referendums: who determines the base price of a vote? If a single vote cost $0.10 or $100, I imagine that could drastically affect outcomes, and I don't see them address this question anywhere.
[1] https://en.wikipedia.org/wiki/Condorcet_paradox
[2] https://poseidon01.ssrn.com/delivery.php?ID=2871141120741160...
It could be tuned to a power of 1.9 or 2.1 or 3 or 1.5.
Aside from being a round number, how do we know 2 is best? What would be the pro/con of a higher exponent? What about the pro/con of a lower exponent?