Friday, July 14, 2017

Why Does Quadratic Voting Excite Miles Kimball?


In yesterday's blog entry I endeavored to encourage Miles Kimball and Robin Hanson to supply a coherent economic story.  Today I learned that Kimball is a fan of quadratic voting (QV).

The first time that I heard of QV was three years ago when memesteading.com asked me about it.  Here was my response...

*******************

Nope, hadn't seen it. Thanks for pointing it out. It's fun to pretend that they got the idea from this blog entry...Crooked Timber Liberals Do Not Advocate Selling Votes.

I just skimmed over the first search result...I'm not sure why 'quadratic'. It seems far more intuitive that the price of votes should be determined by the demand.

And it's a bit off to use quadratic voting or regular vote selling rather than tax choice when it comes to something like a public park.

For me vote selling would be relevant for things like gay marriage. If somebody is completely neutral on the topic...then they would simply sell their vote to the highest bidder.

The 'quadratic vote' people are on the right track though...the answers/outcomes are far more accurate when people are given the opportunity to put their own money where their mouths are (deep input).

*******************

Since then I've read quite a bit about QV, but I still have the same issues.  Perhaps Kimball can address them.

Why quadratic?  Here's an explanation...

Simple vote trading won’t work, because buying a single vote is too cheap and thus a liquid buyer could accumulate too much political power. - Tyler Cowen, My thoughts on quadratic voting and politics as education

Let's take prostitution for example.  I think it should be legal, Kimball disagrees.  I offer to buy his vote.  Should we assume that he will sell it too cheap?  Why in the world would we assume this?  Why would we assume that he only weakly cares about prostitution being illegal?  The very point and purpose of a vote market would be to figure out the (opportunity) cost of a vote, yet Cowen's argument against the vote market is that a vote will be too cheap.  Ouch, my brain.

What would count as "cheap" anyways?  A penny sure is cheap, but what about a quarter, or a dollar?  We'll say that Kimball decides to sell his vote for a quarter.  From my perspective, I suppose this would be pretty cheap.  Well, unless I was trying to buy a million votes for a quarter each.

If Kimball was willing to sell his vote for a quarter, then this would provide some evidence that he didn't strongly care about prostitution being illegal.  Or, maybe he was super broke and could really use the quarter more than he could use the vote.  In either case, what essentially occurred is that he exchanged some power in the public sector for some power in the private sector.  I did the opposite.  I gave up some influence in the private sector in order to gain some influence in the public sector.  To be clear, it's entirely possible that Kimball will use the quarter that I gave him to buy a vote on an issue that he cares more strongly about.

With all of this in mind, what's the point of the quadratic aspect?  What's the benefit of making it progressively more costly to buy additional votes?

If we eliminate the quadratic aspect, then the next question is, why do votes need to be involved?  Kimball and I essentially made a mutually beneficial trade.  But we really didn't need a vote to be part of the trade.

Here's how it would work without votes.  Participants would have a window of opportunity to spend as much of their money as they wanted on their preferred outcome.  After the window of opportunity closed, the totals would be calculated and revealed.  Whichever side had spent the most money would get their preferred outcome.  Not only would the other side receive a complete refund, but they would also proportionally receive all the money spent by the winning side.  It would be a mutually beneficial trade.  Votes are clearly unnecessary.  This is coasianism.

Let's consider a small, real-life, example of coasianism.  The students in my friend's 4th grade class needed to decide who should be in charge of their Justice Department.  Rather than decide by voting, they decided by spending.  Each student wrote three things on a piece of paper...

1. Their name
2. Their willingness to pay (WTP)
3. Their preferred candidate

After they had all turned in their papers, the teacher wrote the WTPs on the board...




Christopher was the winner, despite the fact that he was only preferred by one student... himself.  Here's a picture of his piggy bank...




Christopher decided that it would be worth it to give up his $4 dollars in order to be in charge of the Justice Department.  Of course he couldn't know beforehand how much money the other students were willing to pay for the other candidates.  For all he knew, he might have lost.  In that case, he would have been compensated.  He won instead which meant that he compensated all the other students.  Christopher essentially made a mutually beneficial trade with nearly 30 other students.

It should be crystal clear that votes really do not need to be part of the trade.  All that matters is that participants have the incentive to use their own money to honestly reveal the true intensity of their preferences.

Quick summary about QV.  Is the quadratic needed?  Nope.  Is the voting needed?  Nope.  All that's needed is the participants deciding how much of their own money they are willing to pay for the things they want.

QV is a hybrid made from democracy and markets.  It's definitely far superior to pure democracy, but its democratic traits make it far inferior to pure markets.

A coasian market should be used to determine whether prostitution should be legal or illegal for a year.  If the coasian market determines that it's more beneficial for prostitution to be illegal for a year, then a buchanian market should be used to determine how much money should be allocated to enforcing the law against prostitution.  The buchanian market would simply involve each and every taxpayer having the option to decide how many of their own tax dollars to allocate to the enforcement of the law.

That's public finance in a very simple and straightforward nutshell.

After I tweeted yesterday's blog entry to Kimball, he replied with the following...


He gave people the opportunity to rank over 100 things according to their desirability.  Here was my reply...


He didn't reply to this tweet, but he did reply to another one...

Xero: It would be revolutionary if cheap-talk surveys were remotely effective at measuring well-being. http://pragmatarianism.blogspot.com/2017/03/my-well-being-depends-on-artichokes.html @mileskimball
Kimball: In our approach, it isn't cheap talk. Local incentive compatibility: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3760035/
Xero: If not cheap-talk then the options can and should be sorted by the amount of money that was actually spent on them. http://classtopia.blogspot.com/p/blog-page_22.html
Kimball: You need a budget constraint, but it doesn't have to be money. Quadratic budget constraint with voting power has good properties.
Kimball: See paper here https://link.springer.com/article/10.1007/s11127-017-0414-3
and video of my talk:
https://bfi.uchicago.edu/video/quadratic-voting-application-normalized-gradient-addition-mechanism

I wasn't willing to spend $40 dollars to read the paper, but I was willing to spend an hour watching the video.  I might be mistaken but it seems like he used QV to order the list of over 100 items.  However, real money wasn't used, tokens were used instead.

Let's consider a real-life example of a different way to order a list.  Earlier this year the Libertarian Party wanted to choose a theme for their convention.  Here a list of the potential themes...

All of Your Freedoms, All of the Time
Am I Being Detained!
Be Me, Be Free
Building Bridges, Not Walls
Empowering the Individual
Free Lives Matter
Future of Freedom
I’m That Libertarian!
Jazzed About Liberty
Liberty Here and Now
Life, Liberty and the Pursuit of Happiness
Make Taxation Theft Again
Pro Choice on Everything
Rise of the Libertarians
Taxation is Theft
The Power of Principle

Rather than choose their convention theme by using a cheap-talk survey, the Libertarian Party decided to use a skin-in-game survey...

$6,327.00 - I’m That Libertarian!
$5,200.00 - Building Bridges, Not Walls
$1,620.00 - Pro Choice on Everything
$1,377.77 - Empowering the Individual
$395.00 - The Power of Principle
$150.00 - Future of Freedom
$135.00 - Life, Liberty and the Pursuit of Happiness
$105.00 - Rise of the Libertarians
$75.00 - Free Lives Matter
$42.00 - Be Me, Be Free
$17.76 - Make Taxation Theft Again
$15.42 - Taxation is Theft
$15.00 - Jazzed About Liberty
$15.00 - All of Your Freedoms, All of the Time
$5.00 - Am I Being Detained!
$5.00 - Liberty Here and Now

The "sacred" was ordered by sacrifice.  We can clearly see the disparity in sanctity.

It's reasonable to guess that the free-rider problem affected the results.  But if this problem had been solved, would "Taxation is Theft" have moved to the top of the list?  I suppose it's logical that skin-in-game surveys will be biased against free-riders.

Why didn't Kimball use a skin-in-game survey to rank the items on his list?  As far as I know, he's never even publicly considered a skin-in-game survey.  Neither has any other economist.  But Kimball and quite a few others have certainly considered QV.

Maybe QV solves problems that I'm overlooking.  But the fact that economists haven't even publicly acknowledged skin-in-game surveys, makes it seem highly probable that they are the ones overlooking the obvious.

No comments:

Post a Comment