All of these options have intrigued people over the years. But they all have the same problem: they tend toward having each election decided by the few people who care the most about it. True, ordinary elections don’t reflect people’s strength of preference; people who care a lot have the same influence as people who care a little. But these alternate ways to collect, transfer, and trade votes all have the opposite problem; most everyone’s preferences may be ignored except for the few extremists who care the very most.

However, a simple yet amazing variation can allow collection, transfer, trading, and selling in the voting process, while having elections tend to be decided by a weighted average of how much each voter cares. This amazing variation is: voting quarks.

Hadrons are basic particles in physics that can fly around and smash into things. They are made out of quarks, but (in our world today) those quarks are always stuck inside hadrons, and can’t fly around by themselves. Quarks influence the world via the hadrons of which they are part.

By analogy, a voting quark is a part of a vote that can’t influence an election by itself; it must be part of a vote particle. And voting quarks must be formed into square arrays in order to make votes. So you can use one quark to make a vote of size one, or four quarks to make a vote of size two, or nine quarks to make a vote of size three, and so on.

The key idea here is that elections are won by which ever side has the most votes, with bigger votes counting for proportionally more; but what voters are given are quarks, not votes. For example, each election each voter might be given four new quarks. If no collection, transfer, or trading were allowed, this would be pretty boring, as the only useful option would be to convert those four quarks into one vote of size two to use in this election. (After which that vote, and those quarks, would be gone.)

But if quark collection was allowed, a voter could choose to instead save all these new quarks for future elections. Or they might use one quark this election to make a vote of size one, and save the other three for future elections. Or if they had collected at least five previous quarks, they might add them to these new four quarks to create a vote of size three to use in this election.

Abilities to transfer or trade quarks would work similarly; you’d move the quarks around as allowed by the rules, and then form votes from the quarks as desired to use in each election. The system might even not directly give voters quarks at all, but only sell them quarks for cash.

The main point is that in a system like this people have an incentive to vote in each election roughly in proportion to their strength of preference on that topic. Which allows elections to produce more economically efficient outcomes. And the wider the scope over which quarks can be moved, the wider the scope over which choices could be more efficient.

This point has been plausibly argued in a paper called “Quadratic Voting” by Steven Lalley and Glen Weyl. (Weyl has a related paper with Eric Posner.) They talk about this in terms of buying votes directly with cash, paying proportional to the square of the votes bought. This is an extreme version that I suspect most people will find hard to swallow, at least as the first change to accept. So I designed the above quark language to show how we might move there gradually, such as perhaps by first allowing collection, then transfer, then trading. And we might slowly increase the number of quarks given per election, to approach a more continuous voting.

Steven Levitt has commented positively on the quadratic voting idea, but Tyler Cowen criticized it for encouraging “intense preferences of minorities”. I find that a rather odd criticism, and agree with Eric Posner’s response.

I do have a concern though: this approach would require us to pay more attention to agenda setting. Once votes or quarks can be moved between elections, then every election not only decides an issue, it also creates resources that be used to decide other elections. So we’d want to try to ensure that issues in elections connected by quark moves are similarly important, or perhaps set relative quark prices between them.

Also, the act of introducing an election on some topic ends up being an implicit tax on the people who expect to win that election. They will have to use up quarks there than they can’t use elsewhere. If the status quo is already in their direction, then people who favor the status quo will regret holding an election on that topic, even if they expect to win. Factions may conspire to hold repeated elections that they expect to repeatedly lose, just to tax other factions.

This isn’t an overwhelming objection. We already must pay substantial attention to agenda setting even under ordinary voting. But this does up the ante a bit. So we should try this stuff out slowly, gradually, testing and observing as we go. And we may need to invent new ways to set agendas. But this looks pretty promising, so let’s get started!

Added 11a: OK, on reflection one only has to worry about the relative importance of elections when voters can collect or transfer quarks between them. If voters can instead only trade quarks between elections, their relative importance will be reflected automatically in the relative prices of quarks traded. Also Eric Posner suggests a general agenda mechanism:

Added 10Jan: Commenters are too hung up on money. Money is only relevant in the most extreme version I mentioned, where quarks are bought with cash. Consider instead the other options to only collect, transfer, or trade quarks.

GD Star Rating
Tagged as: ,
• Jeff

In a “Quadratic Voting” scheme, wouldn’t those rich in votes just pay those poor in votes to vote on their behalf (or buy votes on their behalf or whatever)?

• Ronfar

/me shrugs

In theory, “vote buying” can happen in just about any election system if you let people make side deals…

• Daniel

So we ignore the possibility of side deals? Isn’t this some “perfectly spherical cow” kind of assumption?

• Dave Lindbergh

Secret ballots more or less kill that type of “vote buying”.

There is no way to tell if the seller voted the way she promised.

• Daniel

It’s not a critical vulnerability, true. But have you heard about chain voting? It regularly happens in my country. Pretty clever. “If an attacker is able to obtain a blank ballot (by theft, counterfeit, or a legitimate absentee ballot) the attacker can then mark the ballot her chosen candidates and convince (through intimidation or bribery) a voter to take the pre-marked ballot to a polling station, exchange it for the blank ballot issued and return the blank ballot to the attacker. “

• Dave Lindbergh

No, I hadn’t heard about that. Fiendishly clever. (Altho I suppose you could argue there’s no harm if the voter really prefers the payment to the vote.)

But there would seem to be trivial ways to thwart that – just a couple off the top of my head:

* Use big sheets of paper for ballots, require that they be un-creased and un-wrinkled when turned in (makes it hard to hide a pre-marked ballot)

* Electronic voting machines

• kgus

Don’t electronic machines print out your vote so can verify its accuracy? There’s your receipt that confirms you voted ‘correctly’ for a payoff.

• Dave Lindbergh

I hope not! That would defeat the purpose of a secret ballot.

I think the usual thing is they print a log of the vote on a piece of paper that stays inside the machine – for hand-checking in case of something suspicious. But that wouldn’t be handed to voters.

I suppose if you really wanted to you could print a barcode or something like that with an encrypted digest of the vote – that way you could confirm accuracy of the count later based on the paper, without revealing how you voted.

• UWIR

“Altho I suppose you could argue there’s no harm if the voter really prefers the payment to the vote.”

Bribery is not harmless.

• Dave Lindbergh

That raises an interesting question – is vote-buying bribery?

Bribery is paying someone to betray a trust. Usually it involves rent-seeking of some kind, where the bribed directs an illegitimate benefit to the briber – illegitimate because it is comes from a 3rd party (not from the bribed – that would be a simple purchase).

If you think voters have an obligation beyond their self-interest (to vote for the common good, morally, ethically, etc.), then vote-buying is bribery.

But if (as many claim) voters have no obligation to vote any way other than in their self-interest, then vote-buying can’t be “bribery”. If the voter prefers the payment to the vote, no trust is betrayed – voting the way she’s been paid to _is_ in her self-interest (or at least she thinks it is).

On a 3rd hand, majoritarian democracy allows 50% + 1 of the population to legitimately steal from and enslave the minority. So if Alice pays me to vote in favor of stealing from Bob to pay Alice, I’m rent-seeking (since the money is paid by Bob, not by me). So that makes it bribery.

• UWIR

Clearly, voters have a legal obligation beyond their self-interest, so unless you’re going to argue against the legitimacy of laws against vote-buying (and, given that those laws were established by democratically elected government, it would be difficult to do so coherently), vote-selling is betraying a trust.

Having someone offer you money to vote a certain way is a rather artificial way of it being in your self-interest. You are getting the money only because you are voting, making this a case of Goodhart’s Law in action. There is a difference in choosing a vote such that the vote itself is in accordance with your utility function, versus choosing a vote such that your vote plus external events, such as payments, are in accordance with your utility function.

Imagine a pot of \$10,000 and 100 people voting to decide how to split it up. If they split it equally, they each get \$100. Now suppose I offer each of these people \$10 to vote to give me all the money. For each person, it is in their self-interest to defect and take the money. Suppose we’re evaluating the payoff matrix for Alice. If no one else defects, then Alice has a choice between getting just \$100, or \$100 plus \$10. If everyone defects, then Alice has a choice between \$10 or nothing. There is the possibility that 50 people will defect, in which case Alice is choosing between a \$100 cooperate payoff versus a \$10 defect payoff, but how likely is that? If we model each other voter as flipping a coin with probability of p coming up “cooperate”, then the probability of 49 defects is maximized at p slightly lower than 0.5. At p = 0.5, the probability of 50 defects out of 99 is about 0.08. So even in the most favorable world, Alice has a expected value of 0.08*\$90 – 0.92*\$10 = -\$2 for cooperating. Since everyone else has the same payoff matrix, everyone will vote for me to get the money, and I’ll walk off with a nice \$9,000 profit.

My understanding is that in many countries, the precincts with the highest vote percentages for the ruling party get the most pork. This, too, I would consider bribery and a corruption of democracy. In such countries, dislodging the ruling party becomes an extremely difficult coordination problem, and to call such countries “democracies” is an abuse of the term.

• IMASBA

If people lack the information that you’re going to bribe everyone than you walk away with \$9000, but if they knew, your plan wouldn’t work. In general you’re right though: of course people can get screwed over in the long term if they sell their votes and it would affect others as well. It’s similar to the dynamics of trade unions, simply protection against divide and conquer attacks.

• UWIR

I just wrote an extensive post explaining why it’s in everyone’s interests to take the money. If you have some counterargument, then present it. Simply disagreeing with my conclusion without addressing my argument is rude.

Also, “then”, not “than”.

• IMASBA

It’s really not that difficult: the briber and the group of voters cannot both make a profit. All the voters could repeat the calculation you performed above and then decide not to take your bribe, they just need that extra piece of information that says you are trying to bribe everybody and that changes things (they have to intuitively understand that Goodhart’s law you mentioned). But this is all getting off-topic: real voters might not be able to calculate their expected rewad or may have to sell even if they know they will lose something in the long term (if you are hungry today you have no choice but to sell for \$10 today even if it means that some \$200 that was promised to be given to you 5 years from now will be reduced to \$100).

• UWIR

“It’s really not that difficult: the briber and the group of voters cannot both make a profit.”

Whether the group is making a profit is irrelevant. Each voter in this model is making a decision based on their own self-interest.

“All the voters could repeat the calculation you performed above and then decide not to take your bribe”

They could, but why would they?

“they just need that extra piece of information that says you are trying to bribe everybody and that changes things”

How?

“(they have to intuitively understand that Goodhart’s law you mentioned).”

Goodhart’s Law doesn’t change the individual incentives.

• Dave Lindbergh

I mostly agree with you – I do think voters have an obligation beyond pure self-interest (a moral obligation, anyway – I see no way to enforce anything legally).

But many, many people don’t agree.

Many people think the whole point of democracy is that people will, and should, vote for their self-interest.

I constantly hear left-of-center people asking “why do poor southerners vote Republican?”; complaining that they’re “voting against their self-interest”.

[Where there self-interest really lies is a separate issue from whether they _ought_ to vote for it.]

Re “to call such countries ‘democracies’ is an abuse of the term” – well my dictionary seems to think they’re democracies.

That doesn’t mean I’m in favor of it.

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

But if (as many claim) voters have no obligation to vote any way other than in their self-interest, then vote-buying can’t be “bribery”. If the voter prefers the payment to the vote, no trust is betrayed – voting the way she’s been paid to _is_ in her self-interest (or at least she thinks it is).

I’m sorry, but this is utter sophistry. Voting based on self-interest is typically justified on the good Kantian (or else utilitarian ground) ground that if everyone voted their self-interest, the overall good would be achieved.

Everyone voting according to how they’re bribed has a quite different effect.

[I think you’re trying to produce a reductio for voting according to self-interest, but you’ll have to justify your hostility to democracy some other way.]

• Dave Lindbergh

[I think you’re trying to produce a reductio for voting according to self-interest, but you’ll have to justify your hostility to democracy some other way.]

Read more carefully. I’m arguing the opposite.

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

The opposite of which. Of course you’ll say you’re not hostile to democracy, but do you deny you’ve tried to argue that self-interest voting is subject to reduction to the absurd because it implies bribery is OK? (If not, you should write more clearly.)

In any event, at least acknowledge that I’ve shown that self-interest voting and bribery have no connection.

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

After I read your response, I am still not sure what you’re arguing the opposite of. [Perhaps you should write more clearly.]

Do you at least admit that your argument that bribery is a corollary of self-interest voting is either a piece of ridiculous sophistry or a sorry mistake in logic?

• Dave Lindbergh

No.

I’m saying that self-interested voting, without any expectation of a moral obligation to consider the effect on others of the policies one votes for, is equivalent to bribery.

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

So it’s an error in logic rather than sophistry. You say they’re equivalent, and while you’re not explicit about the respect in which they’re equivalent, what you have in mind is that they are equivalent responses to self interest: they equally gratify the voter’s self-interest.

But to say that voters should vote their self-interest is not to say that they should follow their self-interest wherever it leads!

Voting according to self-interest is justified as a practice only in terms of its overall consequences. It’s a way of summing societal preferences. Whereas voting according to who bribes you sums the preferences of only the bribers.

• UWIR

“There is no way to tell if the seller voted the way she promised.”

A lot of criminal activity requires trust. And yet it still happens.

• Dave Lindbergh

It happens only where there is some way of telling if the other side is trustworthy – either repeated interactions, or reputation.

Neither applies here.

• UWIR

I disagree.

• Dave Lindbergh

Care to expand on that disagreement?

Can you give an example of criminal activity which requires trust, but happens routinely without a way to tell if the counterparty is trustworthy (repeated interactions or reputation)?

• UWIR

Do you have evidence for your claim?

My understanding is that much of drug sales, prostitution, bribery, etc. happens based largely on trust. Have you seen the movie Go? Argument From Fictional Evidence, I suppose, but I don’t think that sort of thing is so unusual.

• IMASBA

There is a relevant difference compared to how you would buy votes in the current system. Right now you have to trust someone to change their vote and to not report you to the authorities. In a trading system you merely have to trust an ally to vote the way they voted before.

• http://overcomingbias.com RobinHanson

All voting systems are changed greatly by allowing direct vote buying. Our current voting system forbids that, mostly successfully, so why couldn’t this variation also succeed?

• IMASBA

Sell your vote on a market, or even buying them directly from the government? Yeah sure, what could possibly go wrong with that? I mean it’s not like there are billionaires out there who have strong politicial agendas that greatly differ from what the average person wants, right?

On to the idea of vote saving. I actually like the idea, I don’t think a tyranny of extremely interested voters is something we have to worry about since they’d still have only one vote per election on average and many people already do not vote. The only place where I could see this go wrong are referendums which are often announced a long time before they occur and focus on a single topic, perhaps extremists could conspire to influence those referendums, or politicians may see what’s going on and decide to have a lot less referendums, either of those outcomes would be bad. Quark voting gives more flexibility and is a good idea in principle, but it might be too difficult to understand for a lot of people. Am I too assume that the idea behind having quarks instead of simply having halves or quarters is that it becomes increasingly more expensive to collect votes (you get only 3 votes with 9 quarks but 1 vote with 1 quark), if so than I applaud the idea but fear it might be too complicated for people?

P.S. If I were an American I’d focus more on removing the first-past-the-post element from your elections as much as possible because that encourage gerrymandering, are less accurate and force a two-party system (which in turn is a cause of money playing a greater role in American elections than it would in a country with say 4 large parties).

• http://overcomingbias.com RobinHanson

Yes the key idea of quarks is that the cost per vote is quadratic in quarks.

• http://don.geddis.org/ Don Geddis

Re: your P.S. It certainly sounds intuitive, but David Deutsch in The Beginning of Infinity argues much the opposite. Namely, that a 2-party system with 1st-past-the-post better serves the REAL function of government. Which is not necessarily to accurately represent the current voter preferences of the group (which, as Arrow showed, are necessarily irrational in any case), but instead: to allow clear experiments in public policy, so that, over time, the nation can get closer to truth and learn better what policies actually work. But to do clear experiments, you need a clear winner who takes all of the power. So that they also get the blame when policies fail. In multi-party parliaments, every policy is always a compromise, so failure can always be rationalized with the no-true-Scotsman defense.

(Gerrymandering is different, but California points the way to that fix, which is to separate the drawing of the boundaries, from the politicians whose careers are affected by the boundaries. Politicians shouldn’t draw their own boundaries, because their personal incentives are very different from the good of the voters.)

• IMASBA

Strange… whenever I follow American politics it seems to be about one party arguing that while they had the president/senate/house at the time of some big mistake the other party had the senate/house/president at the same time so the first “party” can’t really be blamed for the mistake (the clear policy experiment argument might work for unicameral parliamentary systems, but then again, having a unicameral system with only two parties is dangerously close to a single-party state). Also (and of course I may be wrong) I’ve come to the conclusion that a two-party, first-past-the-post system increases the influence of money in politics.

• http://don.geddis.org/ Don Geddis

Yes, of course a party in power will try to blame the opposition when there is failure. The question is about the relative success of these political strategies, under different forms of democracy.

Re: money influences in elections. Evidence seems to suggest that this is more perception than reality.

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

No party is “in power” in America today; this would seem to demonstrate that the system doesn’t foster experimentation where one party is in control.

The American system of checks and balances serves just the opposite purpose. It tends to prevent any faction’s idea’s from being tested.

However, the first-to-the-post issue is independent of the parliamentary-versus-American-system issue. It is first to the post that militates toward a two-party system. Deutsch makes the interesting claim (as you summarize him) that the two-party system fosters social experiments. But where we’re limited to two alternatives, parties that represent coherent social visions are impossible. The best experimental procedures don’t ameliorate the careless choice and formulation of hypotheses.

• sanchk

How would foresee quark usage be tracked? One potential method I see would be to implement an electronic voting software which tracks the accumulation and distribution of quarks per election, tied to each voter. You would also want to provide voters transparency over how many quarks they’ve used, which quarks they currently possess and how many they’ll have in future elections to help inform their decision making, which means the database would have to be readable by the general public.

I also wonder if quarks would be accumulated by age (+4 quarks every 2 years, say), or by number of elections which someone is eligible for (so +1 quark for every general election, run-off election, or special election which you are eligible to cast a vote). If accumulated by age, then wouldn’t there be a skew towards elderly voters? But if was by election eligibility, that would present a skew towards districts with higher numbers of elected officials (or which allow run-off elections), since voters would be able to acquire more quarks per cycle of elections. Although, I guess if the additional quarks were of any value to a voter then you would see movement of citizens towards those districts that distribute more quarks…

• IMASBA

Civilized countries have the same number of elected positions per election even in every administrative unit (that is if referendums are treated separately, which there already was good reason to if you want to give people the option to save votes).

• UWIR

So, any country without that property is “uncivilized”? Do you really think this sort of comment is appropriate?

• IMASBA

It was a light-hearted comment meant to point out that the US is fairly unique in having different numbers of elections (not counting referendums) per administrative unit. People are way too Amerocentric on this blog: dealing with having different numbers of elections in different areas is not a fundamental problem of democracy, having different numbers of elections is itself the quirk without which democracy survives just fine in many other countries.

• UWIR

Pointing out that there is a difficulty that would have to be overcome is not “Amerocentric”. Sanchk never said that democracy depends on this “quirk”.

• Dean Jens

The “quadratic voting” with votes purchased for (e.g.) money creates a lot of value for strategic information, even in high-information environments; that is, even if everyone knows approximately how the vote is going to go, having incrementally better information about how close it is going to be helps your side quite a bit. In a more realistic situation, strategic information will be important, but it seems like heuristics and guessing as to how close it is likely to be (or rather how likely it is to be close) will play a bigger role than information. What I like about trading votes for votes rather than for money is that I think there are more natural focal points, so that the outcome of the election will depend less on whether one side tended to overestimate the closeness of the election relative to the other side’s estimate, and more on people making trade-offs between elections purely in terms of the relative importance they place on them.

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

The proposal was developed to allow vote buying. Thus, it aims at a oligarchy (or at least the formal privileging of the rich).

Are we to assume that this proposal, so attractive to you, becomes something qualitatively different because you change a feature to make it more acceptable? Are we to ignore that the proposal might be a preamble to vote buying? [Of course. Conservatives are interested only in the very short term or the very long term.]

The most savvy and those with the best resources will master and dominate such a system.

• stevesailer

Right.

The right side of the Bell Curve has been waging war on the left side for decades.

• charlie

If the function they chose allowed a single vote to be bought for \$10 million, would that strike the right balance for you? Or is this a matter of principle. They claim the quadratic function coincidentally has mathematical properties that strike the right balance. And it does seem fairly punitive for vote buyers: 20 squared = 400, for example.

• stevesailer

Once again: what’s the point? Are billionaires really lacking in political influence at present? Are affluent minority groups like gays and Jews in dire need of more political power?

• SuikaSuika

Flipping it around, you get a huge amount of wealth transferred from the rich to the poor. If quarks are sold at say \$20k each, in order to get a hundred votes, a rich person would need to pay \$200million to a ten thousand people. And a hundred votes isn’t even very significant. It’s a very inefficient return if you’re talking about political influence, as 200 million represents a significant amount of lobbying. So I don’t think this is easily abusable the way you think it is.

• stevesailer

It sounds like First Century BC Republican Rome. How’d that work out?

• UWIR

If the only source of votes is buying them, then the price of them is irrelevant. The dollar/vote ratio is relevant only if there’s some allotment of free votes.

• charlie

The “point” or premise is that one-person-one-vote is an inefficient decision procedure when intensity of preferences varies across voters.

Imagine your town has 2 referenda: 1) should the Klan host their meeting on Steve Sailer’s lawn next Tuesday? 2) Should the Joneses have cherry or apply pie next Tuesday?

Do you think the welfare maximizing procedure is for you and the Joneses to have 1 vote for each question?

• Pshrnk

How might they feel if it were CUBIC?

• Pingback: Assorted links | Homines Economici

• Frederic Bush

I think elections have to be held on a predetermined schedule. Otherwise the people in charge of calling the votes can either decide to buy up all the outstanding quarks now and hold a snap election on a crucial issue, or dilute the value of the outstanding quarks by holding a lot of inconsequential votes before holding an important election. If they can predict in advance which way quark value will move then the group that calls the elections has an enormous edge.

• stevesailer

It’s 2015, not 1975. Do the rich really need more political privileges today?

• http://overcomingbias.com RobinHanson

How exactly does letting everyone shift and trade votes to focus on the issues they care most about favor the rich?

• stevesailer

Well, the rich have more money.

• http://overcomingbias.com RobinHanson

Money is irrelevant except in the most extreme version that I mentioned.

• kgus

Good heavens, man, money is never irrelevant! The strategic disparity between the well-endowed and the rest of us would probably increase (see lump1 and cahokia’s exchange just above).
If one dismisses the wisdom of crowds when it comes to voting, then sure — you’ll have another, thicker, layer of political fog.

• stevesailer

Right. Money hires you MBAs with spreadsheets to figure out how best to exploit this novel system.

• Robert Koslover

Um, isn’t it already difficult enough to ensure basic integrity (lack of cheating) in the voting process? Who are you going to put in charge of running/operating such complex voting systems? The IRS? A computer program? Sure. What could possibly go wrong?

• Pshrnk

Mt. Gox will keep track of our quarks for us.

• lump1

I haven’t read the technical literature, but my first thought is that this would be used to disenfranchise minorities by tying up their quarks on deliberately provocative red herring votes.

For example, let’s imagine a gubernatorial candidate in Texas wants to make sure that Latinos stay away from his election. He and his party arrange a concurrent ballot measure (or some other heinous red herring vote) proposing the right of the police to warrantlessly do a house search whenever they suspect non-compliance with immigration law. Of course the latino minority would pool their votes to reject this stupid measure, but then find that they don’t have any votes left over for the gubernatorial race. In essence they get shut of out the election by gaming the voting agenda.

In general, I feel like a voting system with more moving parts is one that leaves more room for Machiavellian exploitation. This is no exception.

• Cahokia

This proposal only makes sense if one of the driving motivations behind it is to disenfranchise undesirable elements.

A quark-based electoral system would be far too complex for someone even in the middle of the bell curve for intelligence to reasonably navigate.

• mcdowella

Representative democracy is one solution to problems with complexity – have the people elect a representative assembly with some form of proportional representation and then use some more complex system of voting (such as quark-based elections, perhaps with tokens instead of real money) to make decisions in that assembly, which might contain a bewildering assortment of small parties. Having a sophisticated electorate like this might also make it easier to change the voting system frequently without confusing people.

• mcdowella

I propose that a standard voting system such as single transferable vote be used to select a leading and an alternate candidate for senior votes, with a saved vote (or possibly randomized voting system) then used to confirm the leader or reject them in favor of the alternate. The idea would be to dissuade ambitious junior politicians from alienating minorities unnecessarily without handing over too much control.

• SuikaSuika

I see a potential problem of vote buying, though not in the normal way. What I would do if I wanted to buy a hundred votes would not be to buy a ten thousand quarks, but to buy 300 quarks, and give them to 100 people who support my agenda, thus doubling their votes. This strategy sidesteps the quadratic scaling that was the intent of quarks.

• http://overcomingbias.com RobinHanson

If you are confident that they actually agree with your agenda, this is how this is supposed to work. It isn’t a bug.

• IMASBA

Are you sure that’s really how it’s supposed to work? Effectively it means using money to give your side more than one vote per person. If you buy the quarks from people who don’t vote or who are so poor they want to sell their quarks rather than vote you are simply buying votes.

• SuikaSuika

I was under the impression that what made a quark based voting system unique was it’s quadratic scaling. If there was no such scaling, it isn’t too different from a normal vote buying system.

Using the strategy I mentioned, you would be spending 3x more compared to a normal vote buying system for same same amount of influence. Given how wealth is distributed, this doesn’t seem enough to overcome the problems of a vote buying system.

Though now that I think about it, it’s probably more than 3x, since the tripled demand would drive up prices of quarks a lot more.

In light of that, it might also be possible to simply pay the vote sellers a price above the market rate and get them to vote for your agenda. While this possible now, a quark based voting system would also remove the social stigma against such vote/quark trading, possibly making it more prevalent.

• IMASBA

Referendums should really be kept separate from normal elections otherwise the quark system, or any vote saving system cannot function.

• UWIR

All that is needed is confidence that the people will act according to their commitments.

• Anonymous

As I see it, this is overly complicated way to say that influence should grow logarithmically with the number of votes given (or quarks or whatever you want to call them) by a single person.

Use the formula: (votes counted from a person) = 1 + ln(votes given by the person). This gives almost exactly the same effect as the “quark system”.

• http://overcomingbias.com RobinHanson

If someone like you that claims to prefer math can’t tell the difference between a log and a square formula, I’d rather give people a lookup table.

• UWIR

I believe Anonymous is having trouble distinguishing between log and sqrt, not log and square.

• http://greyenlightenment.com/ greyenlightenment

Democracy is not the solution to good policy, so discussion about quarks and votes becomes immaterial in this context

• Cahokia

Better imperfect democracy, even under Hanson’s far-fetched proposal, than a grey enlightened despotism.

• IMASBA

Democracy is not about always having the best policy, it’s about securing orderly and popular transitions of power and having better policy on average.

An enlightened despot brings the best policy, until he gets stabbed in the back or dies of old age and then it’s back to the game of thrones…

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

Who would you consider to have been an enlightened despot?

• UWIR

This basically means that every single candidate is on one ballot. Governor, senator, president, etc. Not only that, but all future governor candidates, every future senator candidates, all future president candidates, etc. If someone runs unopposed, voters will have no reason to vote for them. I guess the candidate would probably vote for themselves just so there’s one vote.

If there’s two candidates, under our system, there’s no reason for anyone to vote for anything but their true preference. It’s only when we get to three candidates that someone might not vote for their preferred candidate. But with this system, even with two candidates, there’s going to be strategic voting. And that means that campaigns will be built around gaming strategic voting. Candidates will try to manipulate public perception of how close the election will be.

There will also be brinkmanship issues: the Republican candidate for president will almost certainly not win the California’s electoral votes. So Democrats might as well spend their quarks on other races. But if none of them vote, then the Republican will win. So Democrats will try to get other Democrats to vote, while not voting themselves, and Republicans will try to get all the Democrats to vote while not voting themselves, or all vote while convincing the Democrats to not vote. These sort of games will be happening in each election. There will be bluffs and counterbluffs: each party will try to coordinate a consistent strategy while trying to keep the opposing party from finding out their strategy (and trying to convince the other party that any reports of their actual strategy are misinformation).

• JW Ogden

I think a better bundling of policy has more potential than any of these schemes.
So we elect a CEO to head up benevolence. He comes with a proposed tax and will run SS, Medicare, SS, Medicaid, SNAP, TANF. We get a CEO of schooling. A CEO of Transportation and infrastructure.
Maybe on state level we hire a company to do healthcare for all in the state. etc.
Then the existing President and congress will only be for defense and law enforcement. They each set their own levels tax.

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

Extreme preferences are usually pathological. Obsessions with gay marriage, the State of Israel, or gun rights (the ones that first come to mind) don’t reflect welfare. These things are not nearly as important to folks as they believe. We might say they are pure signaling-related rather than welfare desiderata.

[Hence I essentially agree with your colleague’s critique.]

• Cahokia

Most people don’t base their politics on maximizing the welfare of Homo sapiens.

Many people prefer to maximize the welfare of their particular nation, religion, or tribe. Others value certain principles above welfare altogether.

Now you can define the above as pathological if you like.

• http://juridicalcoherence.blogspot.com/ Stephen Diamond

What’s pathological is relative to the cultural level of a society. The point isn’t just that the goals are narrow. What creates a pathology is a gross misjudgment of the personal importance of an issue.

• Pshrnk

If the governments sells quarks we can have the rich pay increased taxes voluntarily. Quarks for the rich and lottery tickets for the poor!

• IMASBA

Don’t you think the rich would want something in return for that money? Perhaps something that is not in the best interest of the poor and middle class?

• Pshrnk

Of course! But selling them quarks gives them votes to use to get what they want in return. That is more open than spending the money on campaign donations, lobbying, junkets etc. And at least the money would go into the public coffers.

• IMASBA

Yes, but what good will that do? They’ll surely make a law that subsidizes rich people right back for the amount they pay for quarks.

• Pshrnk

Why stop there? Buy enough votes once to pass a law saying that only people with net worth over \$10 Million can vote, then pass a law abolishing quarks!

• Pingback: Kiezen in het kwadraat - Sargasso