The Economist - USA (2021-12-18)

18 Holiday specials The Economist December 18th 2021

author Eric Posner argue that the method could poten­

tially work well in organisations large and small, from

the  United  Nations  to  presidential  elections,  from

shareholder meetings to homeowner associations. By

their reckoning, it would also have left Mr Trump trail­

ing Mr Kasich in 2016. Vitalik Buterin, co­founder of a

blockchain platform called Ethereum, has used simi­

lar principles to govern his crypto­community’s deci­

sion­making. He describes it as “an initial foray into a

fundamentally new class of social technology”.

In  principle,  this  social  technology  gives  people  a

reason  to  express,  but  not  overstate,  the  intensity  of

their feelings. To see how that works, imagine that one

voter would happily part with $1,000 if it would magi­

cally  guarantee  their  candidate  a  victory.  A  second

more passionate supporter would pay twice as much.

They both calculate that casting an extra vote will in­

crease the odds of victory by the same small percent­

age (say 1%). And they privately value each voice credit

at the equivalent of about $1 (even if they cannot liter­

ally sell them for money).

The first voter enters the poll. Since a vote increases

the chance of their candidate winning by about 1%, it is

worth roughly $10 in their mind (1% of $1000). The first

vote, which costs only one credit or the equivalent of

$1, is a bargain. So is the second, third and fourth. By

the time they have cast five votes (for a combined cost

of 25 credits), a sixth would cost 11 credits (six squared

is 36, which is 11 more than 25). That is the equivalent of

$11. So it is not worth buying.

The more passionate supporter values each vote at

$20. Once they have cast ten votes (for 100 credits), an

eleventh would cost them an additional 21 credits. So

they limit themselves to ten. That is twice as many as

the  first  supporter,  accurately  reflecting  the  fact  that

they care twice as much about the issue.

The result could be quite different under a simpler

system of one­credit­one­vote. If each vote cost a flat

one credit, the passionate voter would snap them up,

until they had exhausted their budget, or until victory

was so assured an additional vote was not worth buy­

ing. The other, milder supporter might do the same, in

which  case  the  result  would  fail  to  differentiate  be­

tween  them.  Or,  knowing  that  stronger  supporters

were  blowing  their  budgets  for  the  cause,  the  milder

supportermightnotbothertobuyanyvotesatall. That

wouldagainmisrepresenttheirtruefeelings.

Distinguishingstrongfrommildsupportisequally

valuableinopinionpolling.In 2016 DavidQuarfoot of

theUniversityofCalifornia,SanDiegoandhis co­au­

thorsputtencontroversialpolicyproposalsto  4,850

Americanvoters.Theyincludedproposalstoraise the

minimum wage, deport illegal immigrants, repeal

Obamacareandtaxtherich.Someofthevoters  were

askedtorespondona conventionalscale(fromstrong­

lyapprovetostronglydisapprove).Othersweregiven a

budgetof 100 voicecreditstospendquadratically. 

Intheconventionalsurvey,peopletendedto gravi­

tatetooneendofthescaleortheother.Theyexpressed

strongapprovalordisapproval,justasonlinereview­

erstendtogivefivestarsornone.Inbothcases,  a

strongstatementcostsnothing.Inthequadratic poll,

peoplefaceda constraint.Expressingvehemence  on

oneissuerequiredthemtoweakentheirstand on an­

other.Thisconstraintforcedthemtobemorediscern­

ingabouttheirpassions.

qv,see

MrWeylsuggestedthemethodtoChrisHansen, who is

nowa statesenatorinColorado.TheDemocratic legis­

latorsusedqvtohelppickwhichofmanypossible

spendingbillstheyshouldpushintheyearahead. It is

alsousedinTaiwantohelpdecideamongtheinnova­

tionscookedupintheannualPresidential“Hackath­

on”,whichchallengescivicentrepreneurstouse data

toimprovepublicservices.InBrazilthecitycouncil of

Gramadohasalsousedquadraticvotingtosetits prior­

itiesfortheyearandtofindconsensusontaxamend­

ments.It isa waytogobeyond“thesimplicityof yes or

no,”saysProfessorDaniel,thecouncil’spresident. “At

firstwefacedsomescepticism,wefounditstrange,”

hesays.Butafterthecouncillorsweretrained,“every­

oneunderstoodtheidea,whichisthesearchfor con­

sensus,fora democracyofthefuture.”

Quadraticvotingdoessoundstrange.Likethe cau­

cus­raceinWonderland,the“bestwaytoexplain it is

todoit”.AndsoTheEconomistisinvitingreaders  to

takepartinanonlinequadraticpoll(seethelink at the

endofthisarticle)modelledontheexperiment by Mr

Quarfootandhisco­authors.Theproposalsinclude:

shouldcountriesimposea carbontax?Shouldthey le­

galiseassisteddying?ShouldBritainreverseBrexit?

Youcanexpresshowyoufeelabouttheseproposals,

andhowstrongly,bytakingpart.

MrPosnerandMrWeylhavebigdreamsfor  their

moderationmachine.Theyimaginea worldin which

citizenswouldspendfroma lifetimebudgetof  voice

creditsonregularreferendums.Anindividualwho felt

stronglyabouta singleissuecouldhoardtheircredits

untilitsmomentarrives,thenblowalloftheir saved

creditsonit.Thatwouldallowpassionateminorities

occasionallytooutvoteindifferentmajorities.

Inthisvision,voterscouldenterandexitelectoral

contestswhenthespiritmovesthem.Inthat  way  it

wouldresembleCarroll’scaucus­race.Andsince even

minoritiescouldoccasionallyprevailoverquadrati­

callyconstrainedmajorities,everyonecouldwinat

somepoint.AstheDodosaid,“Allmusthaveprizes.”n

Even

minorities

could

occasionally

prevail over

quadratically

constrained

majorities

To take part in The Economist's quadratic poll,

visit Economist.com /QuadraticPoll