The Econmist - USA (2021-10-30)

52 United States TheEconomistOctober30th 2021

alyst who worked for Mr Obama. Mr Shor’s

proposed  strategy  is  to  decrease  “educa­

tion  polarisation”—the  gap  between  the

votes  Democrats  get  from  college­educat­

ed Americans and those they get from peo­

ple with no college education—to the level

of  2012.  That  was  the  most  recent  year  in

which whites without college degrees gave

Democrats 40% of their votes or more, ac­

cording  to  Catalist,  a  political­data  firm.

Last year just 37% voted for Mr Biden. Since

only  36%  of  Americans  over  the  age  of  25

have  a  college  degree,  relying  on  them  to

fill  the  gap  is  not  viable.  Whites  without

degrees  make  up  the  largest  share  of  the

population in rural states. 

Mr Shor’s solution is “popularism”: the

idea  that  Democrats  need  to  emphasise

their  most  popular  policies,  such  as  eco­

nomic  redistribution  and  lower  health­

care costs, and avoid topics such as immi­

gration  and  defunding  the  police  that

alienate  working­class  white  voters.

Though “don’t do unpopular stuff” sounds

a glaringly obvious strategy, the debate ov­

er the theory has taken hold of the party.

How  Democrats  should  carry  out  Mr

Shor’s plan is unclear. Education polarisa­

tion  is  not  a  new  trend.  A  study  of  survey

data  stretching  back  to  1948  by  Amory

Gethin,  Clara  Martínez­Toledano  and

Thomas Piketty, all economists, illustrated

parallel trends in many Western democra­

cies. The authors write that it will be hard

to  reverse  unless  Democrats  become  the

champions  of  conservative  positions  on

“sociocultural” issues, such as law and or­

der, the environment and immigration.

Persuading  someone  to  change  their

vote  grows  harder  as  political  identities

deepen. In 2020 the link between a voter’s

partisan identity and self­described ideol­

ogy was the strongest on record. According

to  an  analysis  of  survey  data  by  Charles

Franklin,  a  political  scientist  and  pollster

in  Wisconsin,  a  person’s  stated  ideology

explained  roughly  57%  of  their  partisan­

ship in 2020, up from 49% in 2016 and 20%

three decades ago. This rise in social sort­

ing helps to explain the decrease from 10%

to under 5% in the share of Americans who

regularly  switch  parties,  according  to  The

Economist’s  analysis  of  survey  data  from

the American National Election Studies, a

poll conducted in election years since 1952.

Much  of  Mr  Trump’s  success  in  2016

stemmed  from  his  understanding  that

policy preferences are fickle: identity­loy­

alty and group conflict are more powerful

levers  to  earn  a  person’s  vote.  Democrats

cling to the theory that their popular policy

agenda will eventually secure them a per­

manent majority. But before support for a

wealth  tax,  say,  can  win  over  non­college

whites,  they  will  first  need  to  feel  that

Democratic  candidates  understand  them

and share their cultural outlook. 

There Democrats face a further hurdle:

theirincreasinglynationalisedbrand.Ac­

cording to Jonathan Rodden, a political

scientist,onereasonleft­leaningparties

worldwidehavebeendoingpoorlywithru­

ralworking­classwhitesisthattheircan­

didates struggle to differentiate them­

selves fromtheparty’s ultra­progressive

electedofficialsfromthebigcities.This

imposestheotherworldlinessofveryleft­

leaningDemocraticpoliticianselsewhere

ontotheparty’scandidatesintheheart­

land.RecentlossesforconservativeDemo­

cratsinMontanaandNorthDakotashow

howvulnerabletheyaretothis.

WhentheRepublicanswereina similar

positioninthe1960s,lockedoutofpower

by the union of progressive northern

Democrats and conservative southern

ones,theymustereda realignment.They

usedthebacklashagainstthecivil­rights

andfeministmovementstowooconserva­

tive non­collegewhites. Butentrenched

identitiesnowmakesucha featharder.It

appearsnearlyimpossibleforDemocrats

towinbacka substantialshare—say,5%—

of theblue­collar northern whitesthey

oncereliedonforpower,especiallyinthe

Senate.Theyarelikelytobedisadvantaged

by America’s geography­based electoral

systemformanyyearstocome.n

T

heresultofVirginia’sraceforgo­

vernoronNovember2ndwill“setthe

tone”fornextyear’smid­termelections.

SosaidTerryMcAuliffe,theformer

Democraticgovernorofthestate,ashe

launcheda get­out­the­votecampaign

lastFriday.Ifso,thepartyisintrouble.

PresidentJoeBidenwonthestatebyten

percentagepointslastyear,butMr

McAuliffeisnowpollingjusttwopoints

aheadofhisRepublicanopponent,

GlennYoungkin,accordingtoa statis­

ticalmodelbuiltbyTheEconomist. That

gapissmallenoughforMrYoungkinto

closeatthelastminute.

Ourmodelaggregatespubliclyre­

leasedpollsandattemptsseveralmodest

correctionstotheunderlyingdata.First,

weremoveanyuniformbiasesfromeach

pollingfirm.Thesemayarisefrommeth­

odsofcollectingorprocessingdata—or,

inrarercases,fromtheideologicalbent

ofthepollinghouse.Wethensubtract

anysystematicdifferencesbetween

pollsterswhotrytoensuretheyhave

representativesharesofDemocrats,

Republicansandindependentsand

thosewhodonosuchweighting.This

helpscontrolforthetendencyfornon­

party­weighteddatatobouncearound

morethanparty­weightedpolls;insome

casesit alsoremovesanextralayerof

bias.Lastly,weusethemodeltofita

trendlinethroughalltheadjusted

points.Thismethodyieldeda closer

predictionfortherecentrecallelection

inCaliforniathanmostotherpublic

averagesofpollingdata.

Iftheelectionwereheldtoday,our

modelsuggestsMrMcAuliffewould

defeatMrYoungkinbyabouttwoper­

centagepoints.Thatshouldnotcomfort

theDemocrats.Ourmodelalsofinds

plentyofuncertaintyinthepollingdata:

it givestheformergovernoronlya

roughlytwo­in­three(67%)chanceof

winningtherace.Evena modestpolling

errorcouldblowtheaggregateoffcourse.

In2017,whenthestate’scurrentDemo­

craticgovernor,RalphNortham,wonthe

office,thepollingaverageinVirginia

underestimatedhismarginbysixper­

centagepoints.

Winorlose,thetonetheelectionsets

forthemid­termsislikelytobebadfor

theDemocrats.ThepartyintheWhite

Houseusuallylosesboththesubsequent

electionforVirginia’sgovernorandseats

inthenationalHouseandSenateelec­

tionstheyearafter.Anda closerelation­

shipexistsbetweentheswingagainstthe

incumbentpartyinthosegubernatorial

contestsandintheHousevotenational­

ly.WithMrBiden’snetapprovalrating

underwater(seeLexington)anda trend

againsttheDemocratsincongressional

generic­ballotpolling,evena five­or

six­pointmarginforMrMcAuliffeon

November2ndwouldportenddefeatin

themid­terms.

Virginia’snextgovernor

Down to the wire

A RLINGTON,VIRGINIA

OurmodelsuggeststheraceisuncomfortablytightforDemocrats

Too close to call

United States, Virginia governor election

Voting intention*, 2021, %

Sources:Pollofpolls;TheEconomist

55

50

45

40

35

Aug Sep Oct

GlennYoungkin(Rep)

*Pollsareadjustedfor systematic dierence by pollster

andwhethertheyweight by partisanship or past vote

Terry McAulie (Dem)