Social Mobility June 2010

8/6/2019 Social Mobility June 2010

1/181

SocialMobilityMyths


2/181


3/181

SocialMobilityMyths

PeterSaunders

Civitas:InstitutefortheStudyofCivilSociety

London


4/181

FirstPublishedJune2010

Civitas2010

55TuftonStreet

LondonSW1P3QL

Civitasisaregisteredcharity(no.1085494)

andacompanylimitedbyguarantee,registeredin

EnglandandWales(no.04023541)

email:[email protected]

Allrightsreserved

ISBN9781906837143

Independence: Civitas: Institute for the Study of Civil Society is a

registerededucationalcharity(No.1085494)andacompanylimitedby

guarantee(No.04023541).Civitas isfinancedfromavarietyofprivate

sourcestoavoidoverrelianceonanysingleorsmallgroupofdonors.

All publications are independently refereed. All the Institutes publications seek to further its objective of promoting the advancement of

learning. The views expressed are those of the authors, not of the

Institute.

Typesetby

Civitas

PrintedinGreatBritainby

CromwellPressGroup

Trowbridge,Wiltshire


5/181

Aimhigh

Forthoughyoumaynotreachthesky

Youwillmostcertainlyreachthemountaintops

Teachers inscription inmyfathersautographbookonthedayhe left

schoolattheageof14inthesummerof1939.

Afterleavingschool,myfatherstartedworkasafactoryoperative.At17hejoinedtheRAF,andafterthewarheenlistedatateachertraining

college.Hewentontocompleteasuccessfulandrewardingcareerasa

sportsandscienceteacher,whichincludedalmost10yearsteachingin

Zambia,UgandaandNamibia.

Thisbookisdedicatedtomyfather,andtoallthoseteachers,pastand

present,whoencourage children to riseabove their circumstances, to

takeadvantageof theopportunities thatareavailable to them,and to

exploittheirtalenttothefull.


6/181


7/181

vii

Contents

Page

Author ix

Acknowledgements x

Introduction 1

1 Absolutemobility:

Howtallaretheladders;howlongarethesnakes? 8

2 Relativemobility:

Whobenefitedmostfromtheexpansionofthemiddleclass? 21

3 Issocialmobilityfalling? 34

4 Whatwouldaperfectmeritocracylooklike? 47

5 IsBritainameritocracy? 67

6 Howrobustaretheresearchfindings? 87

7 Policyresponses:Fainthopes,falsestartsandredherrings 101

Appendix:Dowereallywanttoliveinameritocracy? 124

Notes 141


8/181


9/181

ix

Author

PeterSaunderswasuntil1999ProfessorofSociologyattheUniversityofSussex,whereheisstillProfessorEmeritus.Hehasalsoheldvisiting

academicpostsatuniversitiesinAustralia,Germany,NewZealandand

theUnitedStates. In1999hemoved toAustralia tobecomeResearch

Manager at the Australian Institute of Family Studies in Melbourne,

and in 2001 he moved to Sydney to take up the position as Social

Research Director at the Centre for Independent Studies. In 2008 he

returnedtotheUKwherehenowworksasanindependentwriterand

consultant.Hisbooks includeANationofHomeOwners;CapitalismASocialAudit;SocialTheoryandtheUrbanQuestion; Introduction toBritish

Politics;Privatisation andPopularCapitalismandAustraliaswelfarehabit

and how to kick it.He recentlypublished two reports for theLondon

based think tank,PolicyExchange: an analysis of the family tax and

benefits system in Britain, and a critique of the governments child

povertytargets.

Moredetailsofhisworkcanbefoundatwww.petersaunders.org.uk.


10/181

x

Acknowledgements

Ioweaspecialdebtofgratitude toDrRodBondat theUniversityof

Sussex with whom I worked closely in the 1990s analysing socialmobilitydata,andwhowasmainlyresponsiblefordevelopingthepath

model outlined in Figure 1 (p. 86). Some of my early work was

supportedbya smallpersonal researchgrant from theEconomicand

SocialResearchCouncil;theCentreforIndependentStudiesinSydney

supportedmewhenIneededtorespondtocriticismsafewyearslater;

andCivitas inLondonhas funded this latest return to the topic. Iam

particularlygrateful toDavidGreen, theDirectorofCivitas,who first

suggested Ihaveanothercrackat these issues,and toClaireDaleyatCivitas,whohasdoneanexcellentjobeditingthemanuscript.Iamalso

grateful forcomments,help,suggestionsorguidanceatvarious times

in thepast fromBobBlackburn,AlanBuckingham, IanDeary,Geoff

Evans,DavidHitchin,GordonMarshall, TrevorNoble,Geoff Payne,

KenPrandy,andPeterShepherd,althoughnoneofthesepeoplebears

anyresponsibilityforwhatfollows.


11/181

Introduction

Almost30yearsago,ProfessorPeterBauer,aneconomistattheLondon

SchoolofEconomics, attackedwhathe termed the Britishobsession

with class. Britain, he said, sees itself as a peculiarly classdivided

nationwheninrealitywearearemarkablyopensociety.Accordingto

Bauer, there is a strong and pervasive myth in Britain that class

divisions are sharper and more enduring than in other western

industrialised countries,and that socialmovementbetween classes is

rareanddifficulttoachieve.Seekingtorefutesuchclaims,heappealed

toanecdotalandsociologicalevidencetotrytoshowthatBritainhasin

realitybeenarelativelyopensocietyforalongtime.1

Backin1981,whenBauerpublishedthisessay,therewasnotalotof

evidence on contemporary rates of social mobility for him to draw

upon. But in the years since then, sociologists and economists have

beenbusy documenting and analysing peoples social and economic

originsanddestinations.Asaresult,wenowknowagreatdealabout

movementuptheladdersanddownthesnakesofoursociety.Wehave

accuratemeasuresofhowmanypeopleborntolowlyparentsendupinhigh status positions, and how many children with privileged

beginnings fail to make the grade. We know how likely it is that

somebodyborn topoorparentswillendupaffluent,or thatsomeone

borntorichparentswillenduppenniless.

Whatissocialmobility?

Socialmobilityreferstothemovementofindividualsfromoneposition

insocietytoanother.Thismovementmaybemeasuredinthecourseof

their own lives (e.g.by comparing their firstjob after leaving school

with thejob theyendupdoing later in theircareer), inwhichcase it

represents intragenerationalmobility;or itmaybeassessedbycom

paring their current position with the position occupied by their

parentswhentheywerethesameage(intergenerationalmobility).

The positions that are compared may be based on occupational

rankings,incomes,education,orsomemeasureofsocialclassorsocioeconomicstatus.Socialmobilityissaidtooccurwhenindividualsmove

fromalowertoahigherposition(upwardmobility),orfromahigher

toalowerone(downwardmobility)ononeofthesescales.

1


12/181

SOCIALMOBILITY

We can also make reasonably reliable claims about why people

succeedor fail.Muchofwhathappens to individuals in the courseof

their lifetimes is down to chance and idiosyncratic circumstances,but

social science can make some generalisations. We can assess whetherhardworkmakesmuchofadifferencetowherepeopleendup;howfar

intelligencedrivespersonalsuccess;whetherformaleducationandpaper

qualifications matter most; or whether the advantagesbequeathedby

onesparentsbeingbornwithasilverspooninyourmoutharewhat

reallycountintheend.

In1996,Iwroteashortbookwhichassessedtheevidenceonsocial

mobility inBritain thatwas thenavailable tous.EntitledUnequalBut

Fair?,2 the book basically supported Bauers claim that Britain is arelativelyopen society inwhichpeoples achievementsmainly reflect

their own efforts and talents. I suggested that,while competition for

high income and high status employment did not take place on a

completely levelplayingfield,modernBritainwasamuchmoreopen

and meritocratic society thanmost commentators seemedwilling to

acknowledge. In particular, the positionswhich individuals achieved

for themselves in modern Britain had a lot more to do with their

abilities and their hard work than with the social advantages or

disadvantagesbestoweduponthembyaccidentofbirth.

Whatisameritocracy?

Thewordwascoinedin1958byMichaelYounginhiswidelyreadand

influentialbook,TheRiseoftheMeritocracy18702033.Youngdetectedin

postwarBritainagrowingemphasisonsocial recruitmentbased,not

on social origins, but on individual talent and motivation, and heanticipated a time when the higher occupational positions in British

society would all be filled by the most able and hardworking

individuals.Hewashimselfsomewhatworriedaboutsuchanoutcome,

forhethoughtameritocraticsocietywouldrunintosevereproblemsas

a result of the disaffection of the less talented and less committed

majoritycongregatingatthebaseofthesocialpyramid.3

Young defined merit as consisting of intelligence and efforttogether.4Inameritocracy,inotherwords,intellectualabilityandhard

workalonedeterminesocialplacement.

2


13/181

INTRODUCTION

InUnequalButFair?,Ishowedthatratesofsocialmobilitywerenot

markedlydifferent fromwhatwewouldexpect to find inaperfectly

meritocratic society, where peoples achieved socioeconomic status

dependedentirelyontheirabilityandhardwork.Ialsodemonstratedthat, to the extent that Britain fell short of the meritocratic ideal,

circumstancesofbirthoperatedmainlytopreventlessable,higherclass

childrenfromfailing,ratherthantostopmoreable,lowerclasspeople

from succeeding. I concluded that, if you are reasonablybright and

motivated, there is little in modern Britain stopping you from

succeedinginlife(evenifsomerichkidsdogetanunfairstart).

When I wrote that book, interest in social mobility was mainly

limited to a small circle of academics, nearly all of whom weresociologists.Whenthebookcameout,mostofthemeitherignoredit,or

attempted to discredit my arguments. Their response was not

unexpected,foracademicsociologyinBritainwas(andstillis)heavily

dominatedby a leftleaning ideology. Only three per cent of British

sociology professors vote Conservative, and nine out of ten of them

describe their political views as hard left or moderate left.5 This

uniformity has produced an ideological orthodoxy about social

mobilitythatstiflesthewholedisciplineinBritain.

Leftwing academics like tobelieve that Britain is a classridden,

unfairsociety inwhichchildrenborn intomodestcircumstanceshave

the odds massively stacked against them. I call thisbelief the SAD

thesis, for it assumes that social advantage and disadvantage

conferredatbirth iswhatshapespeoplesdestinies.TheSAD thesis is

directly opposed to themeritocracy thesis,which suggests that even

childrenbornintothehumblestofcircumstancescansucceediftheyare

bright and they work hard. The leftwing sociological establishment

regardsanysuggestion thatoccupationalselection inBritainmightbe

takingplace onbroadlymeritocraticprinciples as literally incredible.

The SAD thesis is their dominant paradigm through which all

evidencegetsfiltered,andagainstwhichallargumentsareevaluated.6

In the decade and a half sinceUnequal But Fair? was published,

interest in social mobility has mushroomed. Economists as well as

sociologistshavebeguntostudyit,andpoliticiansaswellasacademicshavestartedtotakeaseriousinterestinit.Unfortunately,however,the

SADthesishascontinuedtodominatemostresearchanddiscussion,be

itacademicorpolitical.

3


14/181

SOCIALMOBILITY

Academicresearchonsocialmobilitydevelopedintoamajorgrowth

industryafter theelectionof theBlairgovernment in1997,and this is

when economists joined sociologists in crunching the numbers.

However, the two disciplines tend to measure mobility in differentways,and thishas led tosomeconfusion.Sociologists like toexamine

movement inandoutofdifferent social classpositions, for they see

class as a powerful indicator of peoples material life chances.7

Economists,ontheotherhand,focusmuchmorenarrowlyonpeoples

incomes, so they measure the number of people who end up in a

different incomeband than that of their parents. As we shall see in

Chapter3, thisdifferenceof focusbetween the two stylesof research

hasgeneratedsomeinteresting,andpuzzling,discrepanciesinfindings.Theelectionof theBlairgovernmentalsomarked thebeginningof

seriouspoliticalinterestinsocialmobility,forNewLabouridentified

theencouragementofsocialmobilityasoneofitscoreobjectives.Very

soon,ministerswere liningup toemphasise their commitment to the

meritocraticideal:8

Thegovernmentsfightisonbehalfofhardearnedmerit,noteasyprizes(DavidMiliband,SchoolsMinisters,2004);

...aBritainwheremerit is thekey tosuccess,where theonly thingthat counts isnotwhereyou come frombutwhatyou are (Tony

Blair,PrimeMinister,2001);

Wemustcreateasocietywhereabilityflowstothetopirrespectiveof an individualsbackground (RuthKelly, Secretary of State for

Education,2005).

Overthelast15years,concernaboutsocialmobilityhasincreasinglycometodominatepublicpolicydebates.Allthreemainpoliticalparties

havepublishedreportsonit,theBlairandBrowngovernmentshaveset

up various inquiries to investigate it, and in November 2008, the

Cabinet Office Strategy Unit announced that: Improving social

mobilityisattheheartofthegovernmentsagenda.9

Manytreeshavebeensacrificed

A number of government reports have been produced on social

mobilityor the lack of it.TheCabinetOffice publishedGettingOn,

GettingAheadinNovember2008.Thisreportedthatsocialmobilityhas

4


15/181

INTRODUCTION

notbeenimproving,anditsetoutaseriesofproposals,mostofwhich

were later incorporated intoaWhitePaper,aimedatcapturingmore

qualityjobsand increasing individualopportunities.Thiswas swiftly

followed inMay2009byAlanMilburns reporton fairaccess to theprofessions, Unleashing Aspiration.10 In his foreword to this report,

Milburn (formerCabinetMinister) claimed that birth,notworth,has

becomemoreandmoreadeterminantofpeoples lifechancesandhe

described Britain as a closedshop society. The report was widely

welcomed in Parliament. Then, in January 2010, Harriet Harmans

NationalEqualityPanelpublished its450page report,AnAnatomy of

Economic Inequality in the UK, which claimed that intergenerational

mobility in theUK is low and that large inequalitiesmake itmoredifficultforpeopletoimprovetheirsituations.11

Not surprisingly, perhaps, this new political enthusiasm for

promoting socialmobilityhasbeenheavily influencedby thebeliefs,

assumptionsandprejudices thatunderpinmostof theacademicwork

done in this area. In particular, the prevailing SAD thesis hasbeen

imported from academe into the public policy realm, and this has

generated four, core socialmobilitymythswhich arenow routinely

repeatedbut rarely, if ever, challenged in public discussions of this

issue:

ThemyththatBritainisaclosedshopsocietyinwhichlifechancesareheavilyshapedbytheclassyouareborninto;12

Themyththatsocialmobility,alreadylimited,isnowgettingworse; Themyth thatdifferencesof abilitybetween individuals eitherdo

notexist,orareirrelevantinexplainingdifferentialratesofsuccess;

The myth that governments can increase social mobility via topdownsocialreengineeringwithintheeducationsystemandforcing

moreincomeredistribution.

Most politicians accept these myths. They assume that social

mobilityinBritainisverylimited,whenitisnot,andthatclassorigins

countmuchmore thanpersonal effortand talent in shapingpeoplesdestinies, when they do not. If anything, these SAD myths have

strengthenedover the last ten to15years,not leastbecause themost

influential group of economists to have joined in this debate has

5


16/181

SOCIALMOBILITY

claimed, not only that mobility is limited,but that it has started to

decline.

This claim has attractedwidespread attention and hasbeen taken

seriouslyby politicians in allparties.On the Labour Party side, oldstyleclasswarriorshaveseizedonapparentevidenceoffallingmobility

rates to justify their belief in rootandbranch, toptobottom social

reform to sweep away traditional class privileges. And the Conser

vativeshavegraspedatapparentevidenceoffallingmobilityratesasa

weapon with which to attack Labours record in office.13 Both sides

havebeenpredisposedtobelievewhattheacademicsaretellingthem.

The result is thatpoliticians from all parties now commonly express

their sense of outrage that a classridden, closed society isbecomingevenmoreclassriddenandevenmoreclosed.

Thisisalldeeplydepressing.Almost15yearsafterpublishingwhat

Ihopedwouldbeanimportantcorrectivetotheprevailing,anddeeply

misleading, sociological orthodoxy about social mobility, this ortho

doxy has not only survived, but has strengthened its claims and

extended itsgripbeyondacademeand into theveryheartofgovern

ment. With all three parties now apparently convinced that bright

childrenbornintohumblebackgroundsarebeingblockedbyanunfair

classsystem from realising theirpotential,and that thingsaregetting

worseratherthanbetter,thescenehasbeensetforyetanotherboutof

expensive (and ultimately fruitless) social engineering aimed at

dismantlingtheimaginedinequitiesoftheBritishclasssystem.

ThereisnoshortageofbigideasaroundWestminsterandWhitehall

abouthowtodothis.Wearebeingtoldthatuniversitiesmustbeforced

to admit children from poorer backgrounds on lower grades while

discriminating against those educated in private schools. Higher

educationmustbeexpandedstill further,and loans forstudents from

poorerbackgroundsmustbesubsidised,orreplacedbygrants.Parents

mustbepreventedfromexercisingchoiceoverwhichstateschooltheir

children attend,with enrolmentsdeterminedby local lotteries.There

mustbeanendtostreamingbyability,morechildrenmustbemadeto

stay at schoolpast theageof 16,andmore taxpayermoneymustbe

spent on vocational training. Early years interventions must beextended and strengthened, and evenmoremoneymustbe redistri

buted to low income families in theneverendingmission to abolish

child poverty. Government must also somehow ensure that the

6


17/181

INTRODUCTION

number of goodjobs (i.e. middle class, nonmanualjobs requiring

degrees and other paper qualifications) continues to grow, so that

workingclasschildrencancontinue tobesiphonedoutof their lowly

backgroundstofillthem.Allthisandmoreiscurrentlybeingproposedinordertoaddressthe

social mobility problem. Yet careful scrutiny of the evidence

demonstrates that this problem either does not exist, or has been

hugelyexaggerated.Beforewecommittospendingmillionsofpounds

wecannotaffordonpoliciestoremoveunfairblockagestotalentedand

hardworkingpeople from lessprivilegedbackgrounds,we shouldat

leastmakesuretheblockagesreallyarethereinthefirstplace.

In the chapters that follow, I have dusted down those parts ofUnequal But Fair? that have stood the test of time, and have

supplemented themwithnewandupdatedmaterialsgathered froma

comprehensive, secondary analysis of the contemporary research

literature. My original intention was simply to produce an updated,

second edition of the book, but so much new material has been

publishedsince1996thatIhaveendedupwithanalmostentirelynew

work which addresses the new political circumstances in which the

socialmobilitydebateisnowbeingplayedout.14

Myaiminpublishingthisnewbookisridiculouslyambitious,forI

hopetoconvinceourpoliticalmasters thatmuchofwhat theybelieve

(orsaytheybelieve)aboutsocialmobilityinthiscountryiseitherfalse

ormorecomplicatedthantheythink.IhopeIhaveabitmoresuccess

withthepoliticianstodaythanIhadwiththesociologicalestablishment

15yearsago.All Iaskof thereader is thatyouapproach thematerial

thatfollowswithanopenandcriticalmind.

7


18/181

1AbsoluteMobility:HowTallaretheLadders;

HowLongaretheSnakes?In thisbook I seek todemonstrate,against the receivedwisdom, that

occupational selection and recruitment in Britain is much more

meritocraticthanmostofusrealiseorcaretobelieve.However,Ienter

threeimportantcaveatsattheoutset.

First, I am not looking at recruitment into the very top (elite)positions in British society. There is no open competition for the

position as the next monarch, nor do you stand much chance of

accumulating the sort of assets enjoyedby theDuke ofWestminster.

EvenentrytoaLabourcabinetseemsaloteasierifyouarethebrother,

son,grandson,daughter,nieceorspouseofaLabourPartyGrandee.1I

do not, therefore, claim that all positions in British society are filled

throughopencompetitiononthebasisofindividualabilityandeffort.

My focus ison the99per centofjobsbelow theelite stratum, rather

thanontheonepercentwhichconstituteit.2

Secondly, most research on social mobility compares the job (or

earnings) that somebody achieves in adulthood with the job (or

earnings)thattheirparentorparentshadatroughlythesameage.Such

analysisobviouslydependsontwogenerationsoffamilieshavingpaid

workthatcanbecompared.Thismeansthatresearchonsocialmobility

has very little to say aboutwhathappens to childrengrowingup injobless, longterm welfaredependent households. Even though their

numbers havebecome increasingly significant over the last 40 years,

thosewithoutjobs(and/orwhoseparentswerewithoutjobs)tendtobe

droppedfromsocialmobilitystudiesaltogether,ortheygetsubsumed

intobiggercategoriesandtheirdistinctivenessislost.

This literature cannot, in consequence, tell us much about

recruitmentinto,andoutof,whathasbeencalledtheunderclass,even

though this is probably where our major problem lies. The socialmobilityliteraturedoesnotgenerallytracethelifetrajectoriesofpeople

whowereabusedorneglectedas children,whoseparentsweredrug

abusers,criminalsoralcoholics,orwhosefathersplayednopartintheir

upbringing.Thesemaywellbe crucial, earlydisadvantages,but they

8


19/181

HOWTALLARETHELADDERS;HOWLONGARETHESNAKES?

are overlooked by the simple social class categories and earnings

bands on which mobility researchers tend to focus. I shall have

something to say at the end of the book about improving the

opportunities forthesechildren,butmostoftheresearchevidenceweshallbelookingatisfocusedonthemainstreamofBritishsociety(the

workingclassandthe middleclass),ratherthanon itsmargins(the

underclass).

Thirdly,evenfocusingonlyonthemainstream,Ishallnotclaimthat

Britain isaperfectmeritocracy.Thevariousadvantagesanddisadvan

tages associatedwith different social class origins do still play some

part in influencing peoples occupational chancessociologists have

notbeen completely wrong or deluded. The point, however, is thatBritain is much more meritocratic than is generally believed. Class

originsarenotveryimportant,norarethesupposedadvantages(such

as private schooling) enjoyed by children born to more affluent or

betterplacedparentsparticularly significant in shapingoutcomes.To

theextentthatitispossibletopredictsomebodysoccupationaldestiny,

itistheirabilityandtheirmotivationthatmattersmuchmorethanthe

social class intowhich theywereborn. InmodernBritain, ifyou are

bright and committed, you are likely to succeed in the occupational

systemirrespectiveofwhatyourparentsdidforaliving,andalthough

thingsarenotperfectandtheplayingfieldisnotcompletelylevel,this

means that our society is relatively open. We are much closer to

achieving a meritocracy than most pundits or politicians seem to

suppose.

TheoriginsoftheSADthesis

TheBritish sociological communityhas long argued that competition

formorefavouredsocialandeconomicpositionsisskewedinfavourof

thoseborn into higher classes, and that the dice are heavily loaded

against anyoneborn into the working class. Because sociologists are

paid towriteabout thissortof thing,andnobodyelse (untilrecently)

had the evidence or the analytical tools necessary to study it, their

viewshavelargelyshapedwhatweallthinkweknowabouttheBritish

classsystem.Downtheyears,sociologistshavearguedthatthemeritocracyideal

is nomore than an elaborate illusion, an ideological confidence trick

designedtowinlegitimacyforasociallyunjustsystemthatperpetuates

9


20/181

SOCIALMOBILITY

middle class privilege. No less an esteemed figure than Professor

Tawney once dismissed the idea of equal competition in Britain as

obviously ajest,3 and the generation of eminent British sociologists

thatfollowedhimwashappytoreinforcehissentiments:Thesocialclasssystem...operates,largely,throughtheinheritanceofproperty,

toensurethateachindividualmaintainsacertainsocialposition,determinedby

hisbirthand irrespectiveofhisparticularabilities...Mostsocialmobilitytakes

place between social levels which are close together... The vast majority of

people still remain in their class of origin... In Britain, allmanner of ancient

institutions and modes of behaviourthe aristocracy, the public schools,

Oxbridge,differences of speech and accent, the relationships of the oldboy

networkfrustratemobility.

(ProfessorTomBottomore,1965)4

Thosewhofearameritocraticsocietyinwhicheveryone,startingmoreorless

equally, would be judged on merit alone, need not therefore be unduly

alarmed: the race is still riggedagainst theworking class competitors... the

upperandmiddleclass...isstilllargelyselfrecruitingandthereforetoamarked

degreesociallycohesive.

(ProfessorRalphMiliband,1969)5

Thechancesthatthosebornintodifferentsocialclasseswillstayinthoseclasses

arestillveryhigh.Thoughthereisagooddealofmobility,mostofitis,infact,

very short range mobility. The myths of long distance mobilityfrom log

cabintoPresidentare,overwhelmingly,mythsasfarasthelifechancesofthe

mass of the population are concerned... it is far too simplistic to describe

modernsocietyaspredominantlyachievementoriented.

(ProfessorPeterWorsley,1970)6

Virtually all movement, whether upward or downward, inter or intra

generational, across the nonmanual/manual division is shortrange... There

hasnotbeenmuchchangeinratesofmobilitybetweenmanualandnonmanuallabourovertheperiodsinceWorldWarI.

(ProfessorAnthonyGiddens,1973)7

Peopleare agooddealmore likely to stay at roughly the same level as their

fathersthantheywouldbeiftherewereperfectmobility...Thosebornintothe

top stratahave enormous advantages in respect of theirjobprospectsas in

respectofsomuchelseoverallothers;notleastinthiscountry.Longdistance

movement especiallyfrombottom to top,aswellas from top tobottomis

uncommon.Mostindividualmobilityisfarmoremodest;andmuchofitstaysononesideor theotherof theconventionaldividing linebetweenwhite and

bluecollarwork... Movement up and down the ladder is inhibitedby sharp

inequalitiesofopportunity.

(ProfessorJohnWestergaardwithHenriettaResler,1975)8

10


21/181


Thisuniformityofopinionmadeahugeimpressiononthediscipline

ofsociologyasitbecameestablishedinBritaininthefirst30yearsafter

the Second World War, and its legacy remains with us to this day.

People likeBottomore,GiddensandWorsleywerecore figures in thedevelopmentofsociologyinthiscountry.Theywerethepeoplewhoset

theagenda,didtheresearchandwrotethetextbooks.Whentheysaid

something, itwaspassedon to subsequentgenerationsof studentsas

receivedtruth.Andalltheseeminentprofessorsassuredusthattherace

inBritainwas rigged,meritocracywasamyth,andwhatdetermined

whereyouendedupinlifewasnotyourabilityoryourdetermination

butthecolourofyourfathersclubtie.

Theseclaimsdidnotrestsolelyonideology.Therewasonepieceofresearchtowhichalltheseprofessorsappealed,astudyoftenthousand

men conductedby Professor David Glass and his colleagues at the

LondonSchoolofEconomics in the late1940s.9Thisstudy found that

social mobility across anything more than a very short range of

occupationalstatuseswasuncommon,and thatmostpeoplestayed in

theclassintowhichtheyhadbeenborn.Theseresultswereconstantly

recycledinthesociologicalliteratureofthe1950s,60sand70sand,they

underpinned theunanimitywithwhich theprofessionsupportedand

promotedtheSADthesis.

Therewere,however, twomajorproblemswithGlasss study.The

firstwasthat,bythe1970s,hisdatawerebadlyoutofdate.Itisinthe

natureofallsocialmobility research thatwearealways lookingback

overour shoulders,athow thingswereagenerationago, rather than

how theyarenow.To compare theachievementsof a current cohort

withthoseoftheirparents,youhavetowaitforthemtogrowupand

establish themselves in careers,by which time 30 or 40 years have

passed.ButtheGlassdatawereevenmoredatedthanthat.Thestudy

hadbeen carried outback in the 1940son a sample ofmen,most of

whomhad leftschoolandstartedwork longbeforetheSecondWorld

Warbrokeout.More than twothirdsof the fathersof thesemenhad

first entered the labourmarketwhenQueenVictoriawas stillon the

throne.

Glasshimselfwarned that thepatternsofmobilityhedocumentedbetween these two generations were unlikely to tell us much about

postwarBritishsociety,forsocialreforms(likethe1944EducationAct)

whichwereexplicitlydesignedtoopenupsocialopportunitieshadhad

11


22/181

SOCIALMOBILITY

little chance to takeeffectby the timehe conductedhis research.But

later generations of sociologists largely ignored these warnings and

treatedGlasss findings as if theywere still applicable to the society

theywerelivingin.Glasssdatawereuncriticallyrecycledforthenext30years,andgenerationsofsociologystudentsweretaughtthatBritain

wasarelativelyclosedsociety inwhich theupperandmiddleclasses

couldsecuretheirpositionforthenextgeneration,andwheretalented

children born into the working class had little realistic prospect of

significant upward movement, even into relatively lowstatus white

collarjobs.10

Thesecond,andevenmoreserious,problemwithGlasssstudywas

that its statistics were badly flawed. Geoff Payne11 calculated that,takingaccountof (a) theexpansion inwhitecollarand contraction in

bluecollarjobsduringthetwentiethcentury,and(b)thehigherfertility

ratesinworkingclassthaninmiddleclassfamiliesduringthisperiod,

thedatareportedbyGlasscouldonlyhavebeenvalidifthenumberof

white collar jobs had declined by 18 per cent in the course of a

generation.Inreality,however,thenumberofwhitecollarjobsincreased

over thisperiodby17percent.Glasss findings, inotherwords,were

impossible given the occupational changes documented by censuses

throughthefirst50yearsofthiscentury.

PaynestopsshortofaccusingGlassofmanufacturinghisdata,but

hedoesnotewithsomefrustrationthereluctanceoflatersociologiststo

treat Glasss findings with the critical scepticism they so clearly

deserved.Hisexplanationforthisappliestodayjustasmuchasitdid

30or40yearsago: Almostwithoutexception,sociologicalwriterson

socialclassinBritainhaveadoptedapoliticalstanceontheLeft.When

Glasswrote that therewas littleupward socialmobility, it reinforced

everything else that theyknew about the class system.Therewasno

incentive, therefore, to challengehis results.12For 30years, from the

1950s to the70s,sociologists inBritainacceptedstatisticswhichcould

not possibly be true because they fitted with their own ideological

prejudices.Thoseprejudicesremainjustasstrongtoday.

The

orthodoxy

is

finally

challenged

Eventually,Glasss conclusionswere challenged. Inwhat remains the

singlemostimportantandrigoroussurveyofsocialmobilityinBritain,

carried out in 1972, Professor John Goldthorpe and colleagues at

12


23/181


Nuffield College, Oxford, finally demonstrated, much to their own

surprise,thatthefundamentalassumptionsofsomuchpostwarBritish

sociology were quite wrong. Based upon a national representative

sampleof ten thousandmen,anddefining socialclassesaccording totheincomeanddegreeofauthorityandautonomytypicallyassociated

with theoccupations theyheld,Goldthorpe found thata lotofpeople

were moving between social classes. This was true comparing the

occupations of sons with those of their fathers (intergenerational

mobility), and comparing peoples first occupations with those

achievedlaterintheircareers(intragenerationalmobility).

Goldthorpessocialclasses

Goldthorpe identifiedseven (andwithsubdivisions,eleven)different

socialclasses,butmostofthesecanbecollapsedintothreemainclasses:

The service class (sometimes also referred to as the salariat), which

consists mainly of professional, managerial and highlevel adminis

trativepositions,aswellastheownersoflargecompanies.Thisservice

classisdividedintohigherandlowergrades(classesIandII);

The intermediateclasses,consistingof routinewhitecollargrades (class

IIIa),personalserviceworkers (IIIb),ownersofsmallscaleenterprises

(thepettybourgoisie,orclassIV),whoare inturndividedintothose

with (IVa) and without (IVb and IVc) employees; and lowergrade

techniciansandsupervisors(classV);

Theworking class, which is made up of skilled (class VI), and semi

skilled and unskilled (class VIIa) manual employees exercising no

supervisoryfunctions,aswellasagriculturalworkers(VIIb).

Thevarious intermediateclassesarenotrankedagainsteachother,so

movement between them does not count as upward or downward

socialmobility.

Analysing movement between his three principal classes, Gold

thorpefoundthatalmosthalfofhissample(49percent)endedupina

social classwhichwas different from that intowhich they hadbeen

born.

Flying in the face of sociological orthodoxy, which insisted that

higherclassparentswereadeptatpassingontheirprivilegedposition

13


24/181

SOCIALMOBILITY

totheirchildren,Goldthorpeshowedthattherewassubstantialmove

mentdownaswellasupthisclasssystem.Healsochallengedthelong

standingbelief thatanymovement thatdoesoccur takesplaceovera

veryshortrange,forhefoundthatlongrangemovementwascommon,includingmovementbothwaysacrossthesupposedwhitecollar/blue

collarbarrier.While59percentofthesonsofserviceclassfathershad

retained their class position, 26 per cent of them had fallen into the

intermediateclasses,and15percenthaddroppedallthewaydownto

theworking class. Similarly,while 57per cent ofworking class sons

hadremainedintheclasstowhichtheyhadbeenborn,27percenthad

entered intermediateclasspositionsand16percenthadrisenintothe

serviceclass.Thinkingthattheseresultsmayhavebeenapeculiarproductofthe

longpostwarboom,andthattheonsetofrecessionfromthemid1970s

might have led to a subsequent diminution in social mobility rates,

Goldthorpewenton in1983 toconducta followup studyusingdata

collected from the general election survey of that year. Again, his

results came as a surprise, for rather than narrowing, the chances of

social movement had expanded in the intervening decade! By 1983,

morethanhalf(53percent)ofthesamplehadchangedclasses,andthe

chancesofworkingclassupwardmovementhadmarkedly improved

(the proportion of working class sons entering the service class had

risen from16percent to22percent,while theproportion remaining

wheretheyhadstartedhadfallenfrom57percentto47percent).

ThisevidenceonwhatGoldthorpe termed absoluteratesofsocial

mobility was enough to convince him that the postwar sociological

orthodoxywaswrong.Whathecalled the closure thesis(whichheld

that top positions are selfrecruiting), the buffer zone thesis (which

claimed thatmovementacross longdistanceswasseverelyrestricted),

and the counterbalance thesis (which suggested thatany increase in

intergenerational social mobility chances had been countered by a

decrease in the opportunities for intragenerationalmovement),were

allflatlyrefutedbyhisfindings.

Later studies by other researchers have confirmed Goldthorpes

findings. At the University of Essex, Gordon Marshall and hiscolleaguesfoundinanationalsurveyinthe1980sthatonethirdofall

men and women in the service class had started life in the manual

working class.13 In Scotland,Geoff Payne found thatmembership of

14


25/181


eventhemostseniorlevelsoftheserviceclasswasextremelyunstable,

onlyaminoritymanagingtoretainsuchpositionsacrossmorethanone

generation, and 14 per cent of service class children ending up in

manual working class jobs.14 And drawing on a series of generalelection surveys which have gathered occupational data, Heath and

Paynefoundthat,throughoutthetwentiethcentury,halformoreofthe

population changed its classposition relative to that of their fathers.

Fewer than half of all menborn to class I fathers stayed in class I

themselves, and four in ten of the sons born to unskilled manual

worker fathers achieved movement into white collar or service class

positions.Upwardmobilitywasextensivethroughoutthecentury,and

becameincreasinglycommonasthemiddleclassexpanded(Table1).

Table1

Absoluteintergenerationalsocialmobilityratesfordifferentbirthcohort

intheUKduringthetwentiethcentury(percentages)15

Birthcohort Pre

1900

1900

1909

1910

1919

1920

1929

1930

1939

1940

1949

1950

1959

Men

Upwardly

mobile 27 29 30 39 38 42 42

Downwardly

mobile 20 21 20 17 18 19 13

Women

Upwardly

mobile 24 22 27 23 29 32 36

Downwardly

mobile 30 32 30 28 27 26 27

Inrecentyears,wehavealsohadaccesstotheresultsofanumberof

nationalpanelsurveyswhichhavefollowedarepresentativesampleof

the population over an extended period of time. One of these, the

National Child Development Study (NCDS), has followed 17,416

individualsfromtheirbirthinoneweekin1958throughtoadulthood.

Thesepeoplewereaged46whentheywerelastquestionedin2004.16Alatercohortof16,571peoplebornin1970(theBritishCohortStudy,or

BCS) has alsobeen tracked, and they too have now reached an age

where their achieved positions canbe compared with those of their

15


26/181

SOCIALMOBILITY

parentsatanequivalentage(theywereaged34whenlastcontactedin

2004).17Inbothcases,awealthofinformationhasbeengatheredabout

classbackgrounds,education,jobhistoriesandpersonalattributes. In

addition, theBritishHouseholdPanelSurvey (BHPS),whichbegan in1991andquestionsthesamesampleofpeopleeveryyear,hasalsobeen

minedfortheinformationitcanrevealaboutsocialmobility.

All of these longitudinal surveys confirm that social mobility in

Britain is extensive and common. Goldthorpe and Jackson18 have

analysedthe1958and1970birthcohortstudiesandreportthatbyage

33,45percentof themenand39per centof thewomen in the1958

birth cohort had moved upwards relative to the social class of their

parents,and27percentofthemenand37percentofthewomenhadmoveddown.For the1970cohort (assessedatage30), theequivalent

figureswere42per cent (men)and41per cent (women)movingup,

and30percent(men)and35percent(women)movingdown.

Table2

Workingclassupwardmobility,andserviceclassdownwardmobility,inthe1958

and1970birthcohorts(percentagesoftotalmenandwomeninsamples)19

Classofdestinationatage30/33

Serviceclass Intermediateclass Workingclass

1958CohortMen:

Serviceclassorigin 18.5 4.7 5.2

Workingclassorigin 15.5 9.9 23.1

Women:


Workingclassorigin 12.9 15.2 23.31970CohortMen:



Women:



Source:JohnGoldthorpeandMichelleJackson, Intergenerationalclassmobility in

contemporaryBritain,BritishJournalofSociology,vol.58,2007,52546.Serviceclass

definedasclasses I, IIand IVa (smallproprietorswithemployees);workingclass

definedasclassesVIandVII.

16


27/181


Theseresultssuggestthatitismoreunusualtodayforsomebodyto

stayinthesocialclassinwhichtheywerebornthantomoveoutofit,

either up or down. Table 2 shows that, inboth the 1958 and 1970

cohorts,moreofthemenandwomenbornintotheworkingclasshadclimbedtoahigherclasspositionthanhadstayedput(forthosebornin

1970, almost twice asmanywereupwardlymobile than remained in

theworkingclass).Movementdownwards,outoftheserviceclass,was

alsofairlycommon.Inbothcohorts,aroundonethirdofthemen,and

almosthalfthewomen,borntoserviceclassparentsfailedtomaintain

theirinitialposition.

Giventhatoccupationalmobilityoftentakesplacebeyondtheearly

thirties (the age at which these panel members career achievementsweremeasured),we can assume that totalmobility (intra aswell as

intergenerationally)willbeevengreater than these figures suggest.20

Clearly,theexperienceofclassmobilityiswidespreadinBritain.

Moreroomatthetop

Looking back, nobody should perhaps have been surprised by the

amount of movement uncovered by Goldthorpe, and confirmed bylatersurveys,forwehaveknownforalongtimethattheoccupational

structure was shifting quite dramatically throughout the twentieth

century.Before1914(whenmostofthefathersofDavidGlassssample

wereenteringwork),aboutthreequartersofallthejobsinBritainwere

inmanualwork(basically,theworkingclassinGoldthorpesschema),

andonlyaquarterwere whitecollar (the intermediateand service

classes).Withthedeclineofmanufacturingemployment,thegrowthof

state sector employment and the impact of technological change onroutinejobs,theseproportionshavenowalmostcompletelyreversed.21

Theshrinkageoftheworkingclassandtheexpansionoftheservice

classoverthelasthundredyearshadtoentailrecruitmentofincreasing

numbers of working class children into the new professional and

managerial positions that werebeing created. There had tobe quite

extensiveupwardsocialmobilityinorderfortheoccupationalchanges

totakeplace.Therewas,asJohnGoldthorpeputsit,moreroomatthe

top,whichmeantpeoplebornintolowersocialstratahadtobesuckedupintotheexpandingmiddleclassesabovethem.

Having said that, however, it is also important to note that

downward mobility has continued tobe significant throughout this

17


28/181

SOCIALMOBILITY

period. InGoldthorpes 1972 survey, four in ten of the sonsborn to

serviceclassfathersendedupinalowerclassthantheirfather,andin

the twobirth cohort studies, theproportion is still around onethird.

Thisiscertainlynotwhatwouldhavebeenpredictedfromtheshiftsintheoccupational structure.With themiddle class expanding, and the

workingclasscontracting,wewouldnotexpecttofindmuchevidence

ofmiddle class children slidingdownwards, yetdownwardmobility

hasbeenquitemarked.

When John Goldthorpe compared social mobility rates across

different countries, he found that downward social mobility was

actually higher in England and Scotland than in any of the other

nations included in his analysis.22 Although critics commonly claimthatBritain is less fluidandopen thanotherEuropeancountries (and

particularly Scandinavia), these data on downward mobility suggest

otherwise,forwithanexpandingmiddleclass,itisarguablytherateof

downwardmobility which tells you most about how open a society

reallyis.

HowdoestheUKcomparewithothercountries?

There is considerable disagreement and confusion in the literature

comparingtheamountofsocialmobility inBritainwith thatrecorded

inotherwestern,developedcountries.Wewoulddowell toheed the

warning of the OECD that any such comparisons shouldbe treated

withagreatdealofcaution.23

Most sociologists thinkmobility ratesare similaracrossmostwestern

countries, and the2001 reviewby theUKgovernmentsPerformance

and InnovationUnit concluded that anydifferences thatdo exist are

modest.24 Economists, however, have started to disagree with this,

and the 2010 report of the National EqualityPanel follows them in

claimingthatmobilityratesinBritainarelowerthaninalmostallother

Europeancountries.Notsurprisingly,thisdramaticclaimhasreceived

alotofmediacoverage.25

The truth is thatmobilitybetween social class positions appears little

different in Britain than in other, comparable nations. The NationalEqualityPanelsowndatashowthatabsoluteratesofupwardmobility

(measured in terms of the proportion ofmen ending up in a higher

class than their fathers) are about the same in Britain as in Ireland,

18


29/181


FranceandGermany,butareabitlowerthanintheNetherlands,Italy

andScandinavia.26Thereportdoesnotdiscussdownwardmobility,but

OECD analysis confirms that we are around the middle of the

internationalrankingsoverall.ThisisborneoutinanalysesbyEriksonandGoldthorpe (whoputEngland8thoutof15nations,withgreater

fluidity than Germany, France and the Netherlands, but behind

Sweden,theUSAandAustralia)andbyBreen.27

TheOECDalsoreportsthatBritainismorefluidthantheaveragewhen

wecompareeducationallevelsattainedbyparentsandtheirchildren.28If

we correlate years of schooling undertakenby children as compared

with theirparents,Britains correlationof 0.31 isweaker than that in

any other developed country except Denmark (i.e. our childrens

education is not stronglypredictedby that of theirparents), andwe

rank6thoutof42nationsonthiscriterionoffluidity.29

The idea that Britain is lagging behind everyone else comes from

economists research on income mobility. An influential group of

economists has found that, when we compare mens earnings with

thoseof theirparents,Britainand theUSAperformpoorlycompared

withCanadaandmostotherEuropeancountries(i.e.thereisastrongerassociationbetweenparentsandchildrensearningsherethaninother

nations).30 The report by the National Equality Panel repeats this

finding. But when we inspect the data more closely, this result is

unconvincing.

Forastart,themostprominentoftheseeconomists,JoBlanden,admits,

There isa lotofuncertaintyabout theUK.Theproblem reflects the

fact thatparental incomesmay changemarkedlyover time (inwhichcase it will matter a great deal at what age parental earnings are

estimated).Todealwiththis,parentalincomescanbeadjustedinthe

lightofotherinformation,suchasparentsoccupationoreducation(so

called instrumental variables). But this produces wildly differing

estimates, and it is not clear which we should accept. Blanden, for

example, reports one correlation between UK sons and parents

incomesof0.44,butnotesthatthisseemsextremelyhigh(eventhough

it has itself been adjusted downwards from an initial 0.58) whencomparedwithanotherofjust0.29.Tocomeupwithherown figure,

Blandenendsupaveragingthetwocontrastingresults,butthisappears

19


30/181

SOCIALMOBILITY

wholly arbitrary, andwehavenowayofknowingwhetherornot it

bringsherclosetoatrueestimate.31

Whenthis(adjusted,averaged,estimated)correlationiscomparedwith

thosefromothercountries,itcomesouthigherthantheotherEuropean

countries examined, and is only exceeded by the USA. This might

indicate that incomemobility inBritain and theUS is low compared

withmost ofEurope,but the StandardErrors on these estimates are

huge(theStandardErrorgivesusthelikelyrangewithinwhichthereal

figureforeachcountrylies).

Because these ranges are so great, they nearly all overlap with each

other,whichmeans thedifferencesbetweenmostcountriesestimatesarenotstatisticallysignificant.Forexample,theUSAranksverypoorly

whileSwedenranksquitewell,butBlandenadmits:Itisimpossibleto

statisticallydistinguishtheestimatesforSwedenandtheUS.Inother

words, they might very well exchange places! She goes on: The

appropriate ranking at the top end is difficult with large standard

errors on the Australian, French, British and US estimatesmaking it

unclearhowthesecountriesshouldberanked.32

Finally,addingevenmoreconfusion,Blandenalsoacknowledges that

anotherstudy,byAndrewsandLeigh,hasestimated incomemobility

for 15 countries (not including Britain) and come up with a very

different set of results. On this calculation, the USA is around the

middleoftherankings,beatingAustraliaandcomingclosetoNorway.

Clearly,whenlookingatincomemobilitystatistics,wearedealingwith

data that contain a lot of error,where international comparisons are

fraught,andwhereadjustmentsandassumptionsmadebyresearcherscanproducealarminglydifferentresults.

Whatarewetoconcludefromallofthis?Britainseemsaboutaverage

on class mobility, quite open on education mobility, and its

internationalrankingon incomemobility is (inBlandensownwords)

unclear.ItisstretchingthedatatothelimitfortheNationalEquality

Panel to conclude from patterns like these that our rates of

intergenerationalmobilityintermsofincomesarelowininternationalterms, and in terms of occupation are below the international

average.33

20


31/181

2RelativeMobility:WhoBenefitedMostfromtheExpansionoftheMiddleClass?

Clearly theBritishclasssystem isa lotmoreopen than the traditional

sociologicalorthodoxyeverimagined.Buttheevidencereviewedinthe

previouschapterhasdonenothing todissuadesociologists from their

view that the British class system is rigid and relatively closed. The

SAD thesis continues todominate conventional thinking in this field,andthereasonisthatresearchersgenerallyprefertofocusonwhatthey

callrelativemobilityratesthanontheabsolutemobilitystatisticsthat

wehavebeenexaminingupuntilnow.

Thefiguresonabsolutemobilityshowalotofmovementgoingon,

but they do not tell us how much of this is due to changes in the

occupational system itself, andhowmuch isdue to greater fluidity

between occupational groups. In technical terminology, statistics on

changing absolute mobility rates cannot distinguish structural

mobilityfrompuremobility.Todothis,wehavetocalculaterelative

mobilityrates,andwhenwedo,manyacademicsbelieveitshowsthat

themiddleclassesstillhavethesystemsewnup,andthattheworking

classisstillbeingunfairlyexcludedfromcompetingonanequalbasis

forthematerialrewardsoursocietyhastooffer.

Wehave seen thathighabsolute ratesofupwardmobilityhave to

somedegreebeendrivenbytheexpansionofthemiddleclassandtheshrinkage of theworking class over the last hundredyears.Mobility

researchersrefertothisasthestructuraleffectupwardmobilityhas

increasedbecausemoreroomhasopenedupatthetopofthesystem.

But what they want to know is who has benefited most from this

increasedroomatthetop?Hastheopeningupoftoppositionsmeant

thatworkingclasschildrennowenjoy thesamesortsofopportunities

tosucceed in lifeasmiddleclasschildren?Orhavethemiddleclasses

also taken advantage of this expansionby securing their childrensfuturesevenmoresuccessfullythantheydidbefore?

Clearly, ifwe are concerned about improvingpeoples life chances,

this question of relative gain is secondary to the question of absolute

gain.Peopleareunlikely to carewhether increasedopportunitieshave

21


32/181

SOCIALMOBILITY

beendeliveredby theexpansionofmiddleclassjobs,ora looseningof

classbarriers.Whatmattersmosttothemisthattheirchancetosucceed

in lifehas improved, irrespectiveofwhether everybody elses chances

have improvedbyasimilardegree.JohnGoldthorpeconfirmed this inhis1972surveywhereheshowedthatmostrespondentswereconcerned,

notwithwhethertheirchildrensopportunitieshadimprovedrelativeto

other peoples children,but with whether their children had abetter

chanceofsucceedingtodaythantheydidinthepast.

In this sense, the most important research findings are those

outlined in the previous chapter. Thebig story is that occupational

mobilityhasbecomecommon,andthatmanymorepeopletodayhave

theopportunitytoachieveamiddleclasslifestylethanwasthecaseinthepast.1

Nevertheless,theattentionofresearchershasbecomefirmlyfixedon

relativities,notabsolutes.Theyacceptthereisnowgreateropportunity

forpeople to achieve amiddle classposition,but the test of fairness

they lay down is that those at the bottom should have availed

themselves of these expanded opportunities to a greater extent than

thoseatthetop.Onlyifthishappenswillrelativemobilityratesshow

anyimprovement.

Measuresofrelativeclassmobility:Disparityratiosandoddsratios

Sociologistsseeking tomeasurerelativemobilityratesgenerallydoso

bycalculatingdisparityratiosandoddsratios.

Disparityratiosexpresstherelativelikelihoodofchildrenfromdifferent

social class origins arriving at the same social class destination. For

example, if 20 per cent of those from working class backgrounds

achievemiddle classpositions,and60per centof those frommiddle

class backgrounds achieve middle class positions, there is a 3:1

disparity ratio in the relative chancesofoccupational successenjoyed

bymiddleclassascomparedwithworkingclasschildren.

Odds ratios are rathermore complicated and are constructed in three

stages. First,we calculate the chances of a child from a higher class

position falling toa lowerclass. If, forexample,50percentofserviceclass children retain theirpositionwhile tenper cent fall all theway

downtotheworkingclass,thisgivesthema5:1chanceofendingupin

theworkingclassasagainstremainingintheserviceclass(eventhough

22


33/181

WHOBENEFITEDMOSTFROMTHEEXPANSIONOFTHEMIDDLECLASS?

thechanceofremainingintheserviceclassiseven).Secondly,asimilar

calculationismaderegardingthechancesofachildfromworkingclass

origins rising all thewayup to a service classposition as compared

with the likelihood of remaining in theworking class. If, say, 60percentofworkingclasschildrenendupinworkingclassjobswhile20per

centmake ittothetop,thiswouldgiveeachworkingclasschilda3:1

chance ofgetting to the top as against remainingwhere they started

from.Thirdly, these two setsofdisparity ratiosare thenexpressed in

relation to each otherbymultiplying the firstby the second. In our

hypotheticalexample,thiswouldproduceafinaloddsratioof15:1(5:1

against service class downward mobility divided by 3:1 against

workingclassupwardmobility).

Middleclass/workingclassdisparityratios

Havingdemolished the conventional sociologicalwisdombydemon

stratinginhis1972surveythatabsolutemobilitywaswidespreadand

extensive, it wasJohn Goldthorpe who went on to rescue the SAD

thesisbyemphasisingthatrelativemobilityratesdidnotseemtohave

changedmuchovertheyears.

Although largenumbersofworkingclassmenhadbeenupwardly

mobile, Goldthorpe found that working class children in his sample

had not improved their chances of occupational success relative to

childrenbornintohighersocialclasses.Everybodyhadgainedmoreor

less equally from the expansion of the middle class during the

twentiethcentury,forwhileitenabledmoreworkingclasschildrento

move up, it also helpedmore middle class children to avoid falling

down. Because the expansion of top positions had been equally

advantageoustothechildrenofallsocialclasses,thegapbetweenthem

intermsoftheirrelativechancesofsuccesswasaswideasever.

Goldthorpeconcludedfromthisthattherewaslittleifanyevidence

ofprogresshavingbeenmade towardsgreateropenness in the class

system, and that no significant reduction in class inequalitieswas in

fact achieved during the postwar period, despite the expansion of

educationalopportunitiesforworkingclasschildrensince1944.2

The relative chances of success or failure faced by the men in

Goldthorpes sample canbegaugedbyanalysing a seriesof disparity

ratioscomparingtheprobabilitiesofchildrenfromdifferentclassorigins

23


34/181

SOCIALMOBILITY

ending up in the same class destination. Goldthorpe reasoned that, if

peoples class origins played no part in shaping their life chances, all

disparity ratios shouldbe 1:1. In other words, a working class child

shouldbejustaslikelyasamiddleclasschildtoendupasadoctororanaccountant,andamiddleclasschildshouldbejustaslikelyasaworking

classchildtoendupasashopassistantorfactoryoperative.

Whatheactuallyfound,however,wasadisparityratioofnearly4:1,

when comparing the chancesof childrenborn to serviceclass fathers

and thoseborn toworkingclass fatherseachachievingaserviceclass

positioninadulthood.Theserviceclasschildren,inotherwords,were

aboutfourtimesmorelikelythantheworkingclasschildrentoendup

in the middle class. A similar disparity ratio of about 4:1 was alsoevidentwhencomparingtheirchancesofendingupinaworkingclass

job,only thistime, thebootwason theother foot,andtheprobability

wasthattheworkingclasschildwouldremaininhisclassoforigin.

Bycomparingdifferentbirthcohortswithinhissample,Goldthorpe

wasabletodemonstratethattheseratioshadremainedfairlyconstant

over time.However, in his followup 1983 survey he found that the

disparity in the chances of service class and working class children

achievinga service classdestinationhadapparently fallenduring the

previousdecadefromaround4:1toaround3:1,althoughthedisparity

intheirchancesofendingupinaworkingclasspositionhadincreased

toalmost5:1.3Thechancesagainstworkingclasssuccesshadtherefore

reduced,whilethoseagainstmiddleclassfailurehadlengthened.

Goldthorpeconcluded fromhisevidencethat therewerepersisting

inequalitiesinchancesofsuccessforchildrenfromdifferentsocialclass

origins. Similar findings were also recordedby the Essex University

teamwhentheyinvestigatedrelativesocialmobilityratesintheir1984

survey. Like Goldthorpe, they too found significant disparities in

relativemobilitychances,andtheyclaimedonthebasisoftheirresults

that, Thepostwarprojectofcreating inBritainamoreopensociety...

hassignallyfailedtosecureitsobjective.4

Itispossibletocalculatedisparityratiosforotherdatasetscollected

since thesestudieswerecompleted. Ihavedone this inTable3which

estimates the relative chances of middle class and working classmalesendingupinmiddleclassorworkingclassoccupations,based

on(a)theHeathandPayneelectionsurveydata(lookingatthewhole

period 196497)5; (b) Breen and Goldthorpes analysis of the two

24


35/181


nationalbirthcohortstudieswhenmemberswereaged23(NCDS)and

26(BCS)6;and(c)myownanalysis(inUnequalButFair?)oftheNCDS

birthcohortatage33.7The threesetsofdatahavebeenanalysedand

classified in slightly different ways and are therefore not strictlycomparable with each other (social class boundaries are treated

differentlyineachcase,andtheagesofrespondentsvary),8andforthe

sakeofsimplicity,womenhavebeenexcludedaltogether(because the

labourmarketisquitestronglygendered,womentendhaveadifferent

occupational profile from men). Nevertheless, taken together with

Goldthorpes earlier (1972) findings, the figures inTable 3 provide a

fairlyclearguidetothescaleofthemobilitydisparitiesexperiencedby

menaround the topandbottomof theBritishclassstructure towardsthe end of the twentieth century (I reserve to the next chapter an

examinationofwhatmayhavebeenhappeningsincethen).

Table3

Disparityratioscomparingmalemiddleclass

andworkingclassoriginsanddestinations

(a)196497electionsurveys:

Relativechanceofbeingin:

ClassI/II ClassV/VI/VII

FatherclassI/II 3.3 setat1

FatherclassV/VI/VII setat1 3.5

(b)NCDS(age23)andBCS(age26):

NCDS:


ClassI/II ClassVI/VII

FatherclassI/II 2.7 setat1FatherclassVI/VII setat1 2.1

BCS:


ClassI/II ClassVI/VII

FatherclassI/II 1.8 setat1

FatherclassVI/VII setat1 2.2

(c)NCDS(age33):


ClassI/II ClassIV/V

FatherOPCSclassI/II 2.6 setat1

FatherOPCSclassIV/V setat1 3.8

25


36/181

SOCIALMOBILITY

LookingatTable3, itisclearthatmenborntomiddleclassfathers

enjoy favourable mobility chances as compared with those born to

workingclassfathers.However,noneofthesedisparityratiosisashigh

as 4:1 (the figure recorded in Goldthorpes 1972 survey), many arebelow3:1,andsomeare less than2:1.Remembering thatwearehere

comparing movement right across the occupational class system

(ignoring shortrange upward and downward mobility of children

reaching the intermediate classes), and that a disparity ratio of 1:1

would indicate no association at all between class origins and

destinations,theseresultsdonotseemtoindicatetheexistenceofhuge

barrierstoupwardordownwardmovement.

Middleclass/workingclassoddsratios

Asecondmeasureofrelativemobilityratesisprovidedbyoddsratios,

and they canbe dramatically large. To calculate an odds ratio, we

basicallymultiplytogethertwodifferentdisparityratios.Thisproduces

a single, summary statistic which is intended to measure overall

fluidityinbothdirectionsbetweenanytwoclasspositions.

Whenwereviewthesociologicalliterature,wefindmuchmoreusebeingmadeofoddsratiosthanofdisparityratios,eventhoughtheyare

much more difficult to interpret substantively (Adam Swift does his

best:Oddsratiosmeasurethestatisticalprobabilityofmembersofone

subgrouphavingsomecharacteristicandnothavinganother,relative

to the statisticalprobability of another group having andnot having

thosesamecharacteristics.)9Oddsratiosarecalculatedmainlyfor the

purposesofstatisticalmodelling,or tocompare overall fluidity rates

overtime,oracrossdifferentsocieties,butstandingaloneasdescriptivestatistics,theyarenotparticularlyilluminating.

This has not, however, stopped researchers from appealing to the

sizeofanoddsratio tosupportasubstantiveclaimofunfairnessand

classprivilege. Inhis1972survey, forexample,Goldthorpecalculated

oddsratiosfordifferentbirthcohortsofserviceclassandworkingclass

menrangingbetween13and19,andwhenhefocusedonthefurthest

extremes,comparingthetopoftheserviceclass(classI,thehigherlevel

professional, managerial and administrativejobs together with largeproprietors)with thebottomof theworkingclass (classVII, thesemi

andunskilledmanualworkers),heendedupwithoddsratiosashigh

as36.Atvariouspointsinhisbook,Goldthorpereferstosuchfiguresas

26


37/181


gross,10andheappealstothemtosupporthisclaimthat,Thereality

of contemporary British society most strikingly and incontrovertibly

deviatesfromtheidealofgenuineopenness.11

LaterstudieshavefollowedGoldthorpesexample.IntheiranalysisofmobilityratesbasedontheUKgeneralelectionsurveys,forexample,

HeathandPaynereportedthat46percentofclassImenstayedinclass

I,whilesixpercentfellallthewaytoclassVII.Againstthis,onlynine

percentofclassVIImenendedup inclass Iwhile38percentstayed

put. This produces an odds ratio of 33:1. They conclude: The odds

ratioswouldallbe1inasocietywheresocialoriginswereunrelatedto

class destinations. The larger the odds ratio, the more unequal the

competition, and the ratio of 33:1 suggests that the competition isextremelyunequal.12

Infact,ittellsusnosuchthing.Itdoesnotfollowfromthefactthat

mobility outcomes are unequal that the competition was unequal

even thoughgenerationsof sociologistshavejumped toprecisely this

conclusion. We have to ask why different people achieve different

outcomes relative to each otherwe cannot simply deduce from the

existenceofanoddsratiogreaterthan1thattheremustbesomesortof

unfairprivilegeorblockageatwork.This isacrucial issueand isone

weshalladdressinmoredetaillaterinthisbook.

Theproblemwithoddsratios

Odds ratioscombinesuccessand failurechances inasinglestatistic.A

significant narrowing of odds ratios therefore requires, not only that

working class chances improve, but also that middle class chances

deterioratethereneeds tobea levellingdownaswellasa levellingup.

TrevorNoblepoints out that,with an expandingmiddle class and a

contractingworking class, it is unlikely that thiswill occur, so odds

ratioswillalmostcertainlyfail toregister the improvementsthathave

occurred in mobility opportunities over time.13 Faced with an

expansionofmiddleclasspositions,itwouldbeextraordinaryifmiddle

classparentsdidnottakeasmuchadvantageofthenewopportunitiesavailableasworkingclassparentsdid.AsNobleputs it: It ishard to

imagine, given the economic, cultural, motivational and other

advantages attributed to amiddle class upbringing, inwhat circum

27


38/181

SOCIALMOBILITY

stanceswideningopportunitiestopursueaninterestingandwellpaid

career would not also be at least equally available to middle class

childrenaswellasanyothers.14

Furthermore,withlessroomatthebottom,thereislessandlesschance

thatpeoplewhoarealreadyatthetopwillfalltothebottom.Weshould,

therefore, expect the improved rates of upward mobility witnessed

during the twentieth century tobe accompaniedby a corresponding

decline inratesofdownwardmobility.But thismeansodds ratioswill

remain moreorless constant, even though opportunities have been

expanding.

BreenandGoldthorpeare irritatedbyarguments likethis.Theyclaimthat with a modicum of sociological imagination it is possible to

conceiveofoddsratiosfallingeveninsituationswherethemiddleclass

is expanding. Ifbarriers toworking classachievementwere to some

degree reduced, they say, then talented working class children will

takeadvantageofanyexpansion inmiddleclassjobs faster thandull

middleclasschildrendo,inwhichcaseoddsratioswillfall.15

ButBreenandGoldthorpearewrong.Even ifworking class childrenseize the lions share of the new opportunities, there is no reason to

believe that odds ratioswill fall.Consider a society of one thousand

individualsdividedintojusttwoclasses,withthreequartersofthemin

theworkingclass,andassumethat theshapeoftheclasssystemdoes

notchange fromgeneration togeneration.Twentypercentofmiddle

classchildren (50 inall)fall into theworkingclass ineachgeneration,

butthisonlycreatesroomforsevenpercentofworkingclasschildren

(50inall)tomoveup,fortheworkingclassissomuchlargerthanthemiddleclass.Measuringsocialfluidityinthissociety,weendupwith

anoddsratioof56(PhaseA).

Now imagine, as Breen and Goldthorpe suggest, that policies are

introducedtoimproveworkingclassmobilitychancestosomedegree.

Indeed, allow these policies tobe so successful that the number of

workingclasschildrenmovingintothemiddleclassdoubles,from50to

100.This,ofcourse,willrequirethatmoremiddleclasschildrenmove

down(forthenumberoftoppositionsremainsconstant),so40percent

ofmiddleclasschildren(100outof250)arenowdownwardlymobile.

Theoddsratioimprovesconsiderably,to9.75(PhaseB1).

28


39/181


But now add Trevor Nobles scenario. The same successful social

reformsare introduced,butat thesame time, theoccupationalsystem

starts changing (Phase B2).Not only do 50 additionalworking class

childrendisplace50childrenfromthemiddleclass,butthenumberofmiddleclasspositionsexpands(say,from250to400).

Socommittedarereformerstoimprovingsocialfluiditythattheytake

advice from Breen and Goldthorpe who tell them odds ratios will

improve if there is a more rapid improvement in the chances of

upwardmobility ofworking class children than of children ofmore

advantagedbackgrounds.Ofthe150newmiddleclassjobs,therefore,

nofewerthan100arereservedforworkingclasschildrenandonly50

gotomiddleclasschildren.Whilemiddleclasschildrenschancesofa

middle class position improveby 25 per cent (from 150 to 200), the

equivalentchancesofworkingclasschildren improveby100percent

(from100to200),amplymeetingBreenandGoldthorpesrequirement.

Thenewodds ratio is then calculatedbut reformers aredismayed to

findthat ithasnowrisento11.Despitethefactthatthesamepositive

discriminationprogrammehasbeen introducedas inPhaseB1, things

haveapparentlygotworseinPhaseB2simplybecausethesizeofthemiddleclasshas increased.This isdespite the fact that thenumberof

working class children experiencingupwardmobilityhasdoubledas

compared with Phase B1, and twice as many of the newly created

middle class positions have gone to working class children than to

middle class children. Social fluidity has nevertheless deteriorated

accordingto theoddsratiosbecausemiddleclassdownwardmobility

rateshaveeasedwiththeincreasingsizeofthatclass.Thisisprecisely

theproblemidentifiedinTrevorNobleswork.

This problem is not limited to hypothetical examplesit occurs in

reality. Goldthorpes own research on his 1972 and 1983 samples

showed that,relative tomiddleclasschildren, thechancesofworking

classchildrenachievingamiddleclassjobshortened4:1 to3:1.Atthe

sametime,however,thechancesofmiddleclasschildrenendingupin

the working class (compared with those of working class children)

lengthened from 4:1 to 5:1. Breen and Goldthorpe interpret thesechangesbyinsistingthatfluidityinthisperiodwasunchangedbecause

relative rates, as measured by odds ratios, were not significantly

differentinthemobilitytablesforthetwodates.16Butthisblindsthem

29


40/181

SOCIALMOBILITY

to the crucial fact that more working class children gained even

though fewermiddleclasschildren lost.Theodds ratio fails topick

thisupbecauseopportunityopenedupforoneclassatthesametimeas

threatswerereducedtotheother.

BreenandGoldthorpemaintainthatthestrengthofoddsratios isthat

theyareinsensitivetothemarginaldistributionsofmobilitytables(a

pointwhichtheyratherpatronisinglyclaimIfailtograsp).Arguably,

however,thisispreciselytheirweakness,forbyfailingtotakeaccount

ofchangesinthemarginaltotalsinmobilitytables(i.e.inthesizeofthe

classesfromandtowhichpeoplearemoving),oddsratiostreatsocial

mobility as if it were a zerosum game. The rules of measurement

ensure that nobody can win unless somebody else loses, which is

preciselywhatNobleisattackingwhenhesuggeststhatpuremobility

cannot in any meaningful way be partialled out from structural

mobility.

All of this has profound implications for public policy debates

(discussed in chapter VII). A policy of improving opportunities for

talented individualsfromallclassestorealize their fullpotentialneed

not requireus to forcemoremiddle class children to fail.Wedonothavetoreduceoddsratiostounity(asGoldthorpesuggests);weneed

only reduce tounity thedisparity ratiomeasuringworking classand

middle class childrens chances of achieving a middle class position

(oncedifferencesofabilityandefforthavebeentakenintoaccount).If

wemeasuresuccessbychangesinoddsratios,however,weshallhave

toensure that thechildrenof themiddleclassesaremade tosufferat

thesametimeasthechildrenoftheworkingclassesgain.Butinreality,

ina fairsociety,bothcangain.Indeed,bothhavebeengainingovera

periodofmanyyears.

Table4showstheoddsratios(calculatedbyGoldthorpeandJackson)17

for mens and womens mobility chances in the two national birth

cohort studies.Mobility for the1958cohort (NCDS) isassessedwhen

theywereaged33,while that for the1970cohort (BCS) isassessedat

age30.Moremobilitywill,ofcourse,haveoccurredsincethen,forwe

notedearlierthatintragenerationalmobilityiscommonandcontinuesthroughoutpeoplesworkinglives,butbytheirearlythirties,wehavea

reasonablygoodideaofthetrajectoryonwhichmanypeopleareset.

30


41/181


Table4

Symmetricaloddsratiosforserviceclass(ClassI&II)andworkingclass

(ClassVI&VII)originsanddestinationsinthe1958(upperfigures)

and1970(lowerfigures)birthcohortstudies

Men:

ClassII* ClassVI ClassVII

ClassI 1.5 7.5 13.8

1.6 10.3 21.8

ClassII* 2.8 9.6

3.0 5.3

ClassVI 1.3

1.6

Women:

ClassII* ClassVI ClassVII**

ClassI 1.2 2.9 5.9

1.4 10.9 11.0

ClassII* 2.3 4.8

4.0 3.6

ClassVI 0.6

1.2

* IncludesClassIVa(selfemployedwithemployees)

**IncludesClassIIIb(personalserviceworkers)

Severalpointscanbenotedfromthistable.Thetinyoddsratiosfor

classes I and II in both surveys, and for both sexes, tell us that

movementbetween thehigherand lower levelsof the serviceclass is

readilypossibleforbothmenandwomen.Thereareclearlynobarriersofanysorthere.Similarly,classesVIandVII(theskilledversussemi

and unskilled working class) appear tobe very open to each other

(oddsratiosnohigherthan1.6).Moreinterestingly,theoddsratiosfor

classesII(lowerserviceclass)andVI(skilledmanualworkers)arealso

surprisingly small (below 3 for both men and women in the 1958

cohort),which suggests there is also extensive interchange (down as

wellasup)betweenthesetwopositions,eventhoughtheyarelocated

atoppositeendsof thesystem. It isonlyreally theextremeendstheupperserviceclassandthesemi andunskilledmanualworkingclass

which throwup largeodds ratioswhen theyarecomparedwitheach

31


42/181

SOCIALMOBILITY

other(rangingfrom6forwomeninthe1958cohort,to22formeninthe

1970birthcohort).

Evenhere, though, the figuresarenotas largeaswerereported in

the earlier surveys.Does this indicate that the class system hasbeengettingmorefluidoverthelast40yearsorso?

Dorelativemobilityratesincreaseovertime?

In his work with Erikson, Goldthorpe famously argued that overall

fluidity (as measured by odds ratios) has tended to remain fairly

constantover time.He foundthiswas true,notonly inBritain,but in

almost all western countries, and he referred to this pattern as

trendlessfluctuation,oraconstantflux.18Heappealedtothisfinding

tochallengetheoristsofmodernisationwhohavelongbelievedthatas

societies industrialise,and theirmarkets freeup,so individualoppor

tunitiesopenuptoo.Goldthorpesuggeststhisisnotthecase,andthat

although absolutemobility has expanded, classprivileges anddisad

vantages continue to make themselves felt in relative rates which

remainfairlyconstant.

There are, however, grounds for doubting this claim. In theiranalysisofelectionsamples,HeathandPaynecalculateaseriesofodds

ratios fordifferentbirthcohorts,and they find theyhavebeen falling

(Table 5). For men and womenbornbefore the Second World War,

odds ratios expressing movementbetween the service class and the

workingclassaresignificantlyhigher than for thoseborn later.Heath

andPaynetest thisapparentchangewithaseriesofstatisticalmodels

whichobligethemtorejectthehypothesisofconstantfluidityformen,

althoughtheresultforwomenislesscertain.InplainEnglish:theoddsratios havebeen getting smaller, which indicates (certainly for men,

andpossiblyforwomen)thatfluidityhasbeenincreasing.

Table5

Oddsratios(serviceclass:workingclass)formenandwomen

borninthefirst60yearsofthetwentiethcentury

Pre1900 1900 1910 1920 1930 1940 195059

Men: 16.0 10.0 19.0 14.0 10.3 5.6 7.7

Women: Na 17.2 15.2 13.4 7.3 10.6 5.8

32


43/181


Other studies confirm this.Analysing theBritishHouseholdPanel

Study,andlookingatpeoplesoccupationsbythetimetheyreached35

yearsofage,Gershuny reports consistently fallingodds ratios (based

onserviceclass/workingclasscomparisons) forgenerationsbornafter1940.19HespeculatesthatGoldthorpemayhavefailedtospotthistrend

because he analysed later cohortswhen theywere younger than the

olderones,whichmeanshecouldhavemissedlatercareermovements.

Drawing on a unique family history data set from the nineteenth

century,which they link tomorerecentsurveydata,Lambertandhis

colleagues also find that fluidity hasbeen increasing over time in

Britain,albeitveryslowly: Wecanseeremarkablyconsistentpatterns

of trend in socialmobility ratesover the last twohundredyears.Thetrendofslowlyincreasingsocialmobilityisparticularlystableformen,

andisalsoclear,thoughlessstable,forwomen.20Theydefinepeoples

class as a point on a continuous, hierarchical scale, and this enables

them to estimate the strengthof the associationbetween fathersand

sonsoccupationsusingcorrelationcoefficients.Theyfindthestrength

ofcorrelationshasbeendecliningover time,and theynotewithsome

alarm the readiness of UK policy makers to accept Goldthorpes

constantfluiditymodelwhenitappearstobewrong.

This isnot theonlyexamplewe shallencounterofpoliticiansand

government advisers developing policies on the basis of faulty

assumptionsaboutsocialmobilitypatterns.

33


44/181

3IsSocialMobilityFalling?

Giventheconclusionofthepreviouschapter,thetitleofthisoneneeds

somejustification.Whywouldanybodybelieve thatsocialmobility is

fallingwhenwehavealreadyestablishedthatithasbeenexpandingfor

abouttwohundredyears?

Theanswer lies in the recent,very influential,workofagroupof

economistswhohavebeen lookingat incomemobility,comparing the1958and1970nationalbirthcohortstudies.Theythinktheyhavefound

a clear reduction in the rate of mobilitybetween these two cohorts.

Childrenbornin1970aresaidtobefacinggreaterobstaclestosuccess

than thosebornjust twelve years earlier. This claim has sent shock

waves through thepoliticalsystem,sparkingenquiries,parliamentary

debates and a spate of policy initiatives intended to sponsor more

working class achievement while weakening the tenacious grip on

successofthechildrenofthebourgeoisie.

Incomemobilityinthe1958birthcohort

Toevaluatetheseeconomistsclaims,weneedtounderstandthatthey

are analysing a different phenomenon from the one we have been

lookingatuntilnow.The sociological literaturewehave reviewed in

the last two chapters is concerned with class mobility, whereas the

economistsarelookingatincomemobility.

Classmobilityversusincomemobility

Sociologists prefer to analyse movement between classes because,

although definitions of class are disputed, and its measurement is

inexact, the ideaof class capturesamuchbroader senseofpeoples

lifechances than simply income.Whetheryoucall them the service

class,thesalariat,orthemiddleclass,highlyqualifiedprofessionals,

togetherwith topmanagers and administrators and successful entre

preneurs,shareanumberofmaterialcircumstancesincommonwhich

tend to separate themquite sharply from themanual working class.

Theytendtoearnmore,buttheyalsoenjoygreatersecurityofemploy

ment, they tend to have better retirement pensions (although

34


45/181

ISSOCIALMOBILITYFALLING?

nowadays, this increasingly depends on public sector employment),

theyaregenerallybettereducated,theyhaveaculturalcapitalaswell

aseconomicassetswhichtheycanpassontotheirchildren,theyenjoy

better lifetime health, they tend to be more highlyregarded in thecommunity (and theyhavehigher selfesteem too), and they exercise

more control over their own work lives, and the lives of those they

employormanage.

Class,moreover,tendstobemoreenduringthanmereincome.Income

canbeephemeralaskwhatsomebody isearning today,and itcould

bequitedifferentthistimenextyear.Classtoo,ofcourse,cananddoes

change(wehaveseenthatintragenerationalmobilityiscommon),but

it offers a more reliable indicator of somebodys longterm life

prospects than one or two simple snapshots of their income can

provide.1

Economists, on the other hand, prefer to work with income data.

Income is, inprinciple,more easilymeasured (although you have to

trust thatpeopleare tellingyou the truthaboutwhat theyearn),and

therearefewerdoctrinaldebatesabouthowtodefineit(althoughitcan

bedifficultknowingwhattoinclude,andcalculatingnetincomesafterreceiptofbenefitsandpaymentoftaxcanbeproblematicwhenrelying

onsurveyanswers).It isasimple indicatorofpeopleswellbeing,and

(unlikeclass)everybodyunderstandswhatyouaretalkingabout.

There are also methodological advantages. Because you are dealing

with real numbers, you can use much stronger statistical tools to

analyse incomedata.Andyoudont encounter the clunkingproblem

(which sociologistshaveneververy satisfactorily resolved)ofhow toclassifyhouseholdswhereonememberofa coupleoccupiesahigher

class than theother (this isbecomingamajorproblem in researchon

socialmobilityasmorewomenhavecareers;whichparentsjobistobe

used todeterminea childs classoforiginwhenbothare in fulltime

employmentbut occupy different class positions?). If you are using

incomedata,thisproblemcanbesidestepped:simplyaddtheparental

incomestogether.

One important implication of the economists decision to study

income rather than class mobility is that the concept of absolute

mobilitydisappearsinthistraditionofwork.Thisisbecausemovement

35


46/181

SOCIALMOBILITY

takesplacebetweendifferent incomebands (usuallyquartiles) so the

marginal totalsare fixed inadvance.There isaone in fourchanceof

endingup inanygivenquartile,and this remains the casenomatter

howmuchtheincomedistributionmightchangeovertime.Themiddleclassmaygetbiggerastheworkingclassshrinks,butincomequartiles

remainproportionatelythesamesizeforever.2

Early workby economists analysing income mobility in the 1958

birthcohortproducedresultsbroadlycomparabletothoseproducedby

sociologistsonrelativeclassmobility.Dividingthesampleintoincome

deciles,InstituteforFiscalStudiesresearchersshowedthat,byage33,

men in the top tenthofearningshad fathersdrawn from rightacross

theincomerange.3Nevertheless,therewasaclearrelationshipbetweenfathersandsonsincomes:21percentofthefathersofmeninthetop

decilewerethemselvesalso inthetopdecile(twiceasmanyaswould

beexpectedby chance),whileonly fiveper centof themwere in the

bottomdecile (half asmany as there shouldhavebeen if incomes of

parentsandtheirchildrenwerenotlinked).Thisequatestoadisparity

ratioofaround4:1.

When economists tried to estimate the strength of this association

between parents and childrens incomes, they found very weak

correlations(r=0.24forsonsandr=0.35fordaughters).However,someof

themsuspectedthiswasbecauseparentsincomeswerebeingmeasured

atjustonepointintime,sotheestimatestheyweregettingwereskewed

by transitory fluctuations. They tried to adjust for thisby estimating

parentslongtermearningsfromotherinformationcollectedaboutthem

atothersweepsofthesurvey,andaftermakingtheseadjustments,they

cameupwithmuch strongercorrelationsofbetween0.4and0.6.They

concluded:Weareconfidentthattheextentofmobilityisverylimitedin

terms of earnings, although they did concede that upward mobility

(among thoseborn toparentsat thebottomof the incomedistribution)

wasnotuncommon.Aswithclassmobility,sotoowithincomes,mostof

thestickinesswasduetotherelativelylowprobabilityofthosebornto

therichestparentsendinguppoor.4

Anoteonmeasuringcorrelation

Statisticians commonly measure the strength of associationbetween

two (interval)variablesbyusingastatisticknownasPearsonsProduct

Moment Correlation Coefficient, whose symbol is r. The Pearson

36


47/181

ISSOCIALMOBILITYFALLING?

correlationcoefficientisanumberbetween 1and+1(anegativevalue

indicates that the relationship is inversei.e. as the value of one

variable increases, thevalueof theotherdecreases).The stronger the

association,thehigherthevalueofthecoefficient(avalueof0thereforeindicatescompleteabsenceofassociation,while+1or 1 indicates that

onecompletelypredictstheother).Correlationslessthanplusorminus

0.3aregenerallyregardedasquiteweak.

It is important toremember thatcorrelationdoesnotprovecausation.

Twovariablesmaybecorrelate

Social Mobility June 2010

Documents

Transcript of Social Mobility June 2010