Social Capital in the Creation

of Human Capital'
JamesS. Coleman
Universityof Chicago

In thispaper, the conceptof social capital is introducedand illus-

trated, its formsare described, the social structuralconditions
underwhichit arises are examined,and it is used in an analysisol
dropoutsfromhigh school. Use of the conceptof social capital is
partof a generaltheoreticalstrategydiscussedin the paper: taking
rationalactionas a startingpointbut rejectingtheextremeindivid-
ualisticpremisesthatoftenaccompanyit. The conceptionof social
capital as a resourcefor action is one way of introducingsocial
structureinto the rational action paradigm. Three formsof so-
cial capitalare examined:obligationsand expectations,information
channels,and social norms.The roleof closurein the social struc-
turein facilitatingthefirstand thirdoftheseformsofsocial capital
is described.An analysisof the effectof the lack of social capital
available to highschool sophomoreson droppingout of schoolbe-
foregraduationis carriedout. The effectofsocial capitalwithinthe
familyand in the communityoutsidethe familyis examined.

There are two broad intellectualstreamsin thedescriptionand explana-

tionof social action. One, characteristic
oftheworkof mostsociologists,
sees theactoras socializedand actionas governedbysocial norms,rules,
and obligations.The principalvirtuesofthisintellectualstreamlie in its
abilityto describeactionin social contextand to explaintheway actionis
shaped, constrained,and redirectedby the social context.
The otherintellectualstream,characteristic oftheworkofmostecono-
mists,sees the actoras havinggoals independently arrivedat, as acting
independently, and as whollyself-interested. Its principalvirtuelies in
havinga principleof action,thatof maximizingutility.This principleof
action,togetherwitha singleempiricalgeneralization (decliningmarginal
utility)has generatedthe extensivegrowthof neoclassicaleconomicthe-

1 I thankMarkGranovetter, SusanShapiro,and ChristopherWinshipforcriticisms

of
shouldbe sentto
inrevision.Requestsforreprints
an earlierdraft,whichaidedgreatly
JamesS. Coleman,Department ofChicago,Chicago,Illinois
ofSociology,University
60637.
? 1988by The University
ofChicago.All rightsreserved.
0002-9602/89/9407-0010$1.50

AJS Volume 94 SupplementS95-S120 S95

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

ory, as well as the growthof politicalphilosophyof several varieties:

utilitarianism,contractarianism,and naturalrights.2
In earlierworks(Coleman 1986a, 1986b),I have argued forand en-
gaged in the developmentof a theoreticalorientationin sociologythat
includescomponentsfromboththeseintellectualstreams.It acceptsthe
principleof rationalor purposiveactionand attemptsto show how that
principle,in conjunctionwithparticularsocial contexts,can accountnot
onlyforthe actionsof individualsin particularcontextsbut also forthe
developmentof social organization.In the presentpaper, I introducea
conceptualtool for use in this theoreticalenterprise:social capital. As
backgroundforintroducingthis concept,it is usefulto see some of the
criticismsof and attemptsto modifythe two intellectualstreams.

CRITICISMS AND REVISIONS

Both these intellectualstreamshave serious defects.The sociological
streamhas what may be a fatalflawas a theoreticalenterprise: theactor
has no "engineof action." The actoris shaped by the environment, but
thereare no internalspringsof action that give the actor a purposeor
direction.The veryconceptionofactionas whollya productoftheenvi-
ronmenthas led sociologiststhemselvesto criticizethis intellectual
stream,as in Dennis Wrong's(1961) "The OversocializedConceptionof
Man in Modern Sociology."
The economicstream,on the otherhand, fliesin theface of empirical
reality:persons'actionsare shaped, redirected,constrainedby the social
context;norms,interpersonal trust,social networks,and social organiza-
tionare importantin thefunctioning notonlyofthesocietybutalso ofthe
economy.
A number of authors from both traditionshave recognizedthese
difficulties
and have attemptedto impartsomeoftheinsightsand orienta-
tions of the one intellectualstreamto the other.In economics,Yoram
Ben-Porath(1980) has developed ideas concerningthe functioningof
what he calls the "F-connection"in exchangesystems.The F-connection
is families,friends,and firms,and Ben-Porath,drawingon literaturein
anthropologyand sociologyas well as economics,shows the way these
formsofsocial organizationaffecteconomicexchange.OliverWilliamson
has, in a numberof publications(e.g., 1975, 1981), examinedthe condi-
tionsunderwhicheconomicactivityis organizedin different institutional
forms,thatis, withinfirmsor in markets.Thereis a wholebodyofwork
in economics,the "new institutional economics,"thatattemptsto show,

2 For adiscussionoftheimportance oftheempiricalgeneralization

to economics,
see
Black, Coats, and Goodwin(1973).

S96

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

withinneoclassical economictheory,both the conditionsunder which

particulareconomicinstitutions arise and the effectsoftheseinstitutions
(i.e., of social organization)on the functioning of the system.
There have been recentattemptsby sociologiststo examinethe way
social organizationaffectsthe functioning of economicactivity.Baker
(1983) has shown how, even in the highlyrationalizedmarketof the
Chicago Options Exchange, relationsamong floortradersdevelop, are
maintained,and affecttheirtrades.More generally,Granovetter(1985)
has engagedin a broad attack on the "undersocializedconceptof man"
thatcharacterizeseconomists'analysisofeconomicactivity.Granovetter
firstcriticizesmuch of the new institutional economicsas crudelyfunc-
tionalistbecause the existenceof an economicinstitutionis oftenex-
plainedmerelyby thefunctionsit performs fortheeconomicsystem.He
arguesthat,even in the new institutional economics,thereis a failureto
recognizethe importanceof concretepersonalrelationsand networksof
relations-what he calls "embeddedness"-in generating trust,in estab-
lishingexpectations,and in creatingand enforcing norms.
Granovetter'sidea of embeddednessmay be seen as an attemptto
introduceinto the analysisof economicsystemssocial organizationand
social relationsnot merelyas a structurethatspringsintoplace to fulfill
an economicfunction,but as a structure withhistoryand continuity that
give it an independenteffecton the functioning of economicsystems.
All thiswork,bothby economistsand bysociologists,has constituted a
revisionistanalysis of the functioning of economicsystems.Broadly,it
can be said to maintainthe conceptionof rationalactionbut to superim-
pose on it social and institutionalorganization-eitherendogenously
generated,as in thefunctionalist explanationsofsomeofthenew institu-
tional economists,or as exogenousfactors,as in the more proximate-
causallyorientedworkof some sociologists.
My aim is somewhatdifferent. It is to importtheeconomists'principle
of rationalactionforuse in the analysisof social systemsproper,includ-
ingbut notlimitedto economicsystems,and to do so withoutdiscarding
social organizationin theprocess.The conceptofsocial capitalis a toolto
aid in this.In thispaper, I introducetheconceptin somegenerality, and
thenexamineits usefulnessin a particularcontext,thatof education.

SOCIAL CAPITAL
Elementsforthesetwo intellectualtraditionscannotbe broughttogether
in a pastiche.It is necessaryto beginwitha conceptuallycoherentframe-
workfromone and introduceelementsfromtheotherwithoutdestroying
thatcoherence.
I see two major deficiencies in earlierworkthatintroduced"exchange

S97

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

theory"into sociology,despitethe pathbreakingcharacterof thiswork.

One was thelimitationto microsocialrelations,whichabandonstheprin-
cipal virtueofeconomictheory,itsabilityto makethemicro-macro tran-
sitionfrompair relationsto system.This was evidentbothin Homans's
(1961) work and in Blau's (1964) work. The otherwas the attemptto
introduceprinciplesin an ad hoc fashion,such as "distributive justice"
(Homans 1964,p. 241) or the"normofreciprocity" (Gouldner1960).The
formerdeficiencylimitsthe theory'susefulness,and the lattercreatesa
pastiche.
If we begin with a theoryof rationalaction,in whicheach actorhas
controlover certain resourcesand interestsin certain resourcesand
events,thensocial capital constitutes a particularkindof resourceavail-
able to an actor.
Social capital is definedby its function.It is not a singleentitybut a
varietyofdifferent entities,withtwo elementsin common:theyall consist
of some aspect of social structures, and theyfacilitatecertainactionsof
actors-whetherpersonsor corporateactors-within thestructure. Like
otherformsof capital, social capital is productive,makingpossiblethe
achievementof certainends that in its absence would not be possible.
Like physicalcapital and human capital, social capital is not completely
fungiblebut may be specificto certainactivities.A givenformof social
capital that is valuable in facilitatingcertainactionsmay be useless or
even harmfulforothers.
Unlikeotherformsof capital, social capital inheresin the structure of
relationsbetweenactorsand amongactors.It is notlodgedeitherin the
actorsthemselvesor in physicalimplementsof production.Because pur-
posive organizationscan be actors ("corporateactors")just as persons
can, relationsamong corporateactors can constitutesocial capital for
themas well (withperhapsthe best-knownexamplebeingthe sharingof
informationthat allows price-fixing in an industry).However, in the
presentpaper, theexamplesand area ofapplicationto whichI willdirect
attentionconcernsocial capital as a resourceforpersons.
BeforeI statemorepreciselywhat social capital consistsof,it is useful
to give severalexamplesthatillustratesome of its different forms.
1. Wholesalediamondmarketsexhibita property thatto an outsideris
remarkable.In the processof negotiatinga sale, a merchantwill hand
over to anothermerchanta bag of stonesforthe latterto examinein
privateat his leisure,with no formalinsurancethat the latterwill not
substituteone or moreinferior stonesor a paste replica.The merchandise
maybe worththousands,or hundredsofthousands,ofdollars.Such free
exchangeof stonesforinspectionis importantto the functioning of this
market.In its absence, the marketwould operatein a muchmorecum-
bersome,muchless efficient fashion.

S98

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

Inspectionshows certainattributesof the social structure.A given

merchantcommunityis ordinarilyveryclose, both in the frequencyof
interaction and in ethnicand familyties. The wholesalediamondmarket
in New York City,forexample,is Jewish,witha highdegreeofintermar-
riage,livingin the same community in Brooklyn,and goingto the same
synagogues.It is essentiallya closed community.
Observationofthewholesalediamondmarketindicatesthattheseclose
ties, throughfamily,community,and religiousaffiliation, providethe
insurancethatis necessaryto facilitatethetransactionsin themarket.If
anymemberofthiscommunity defectedthroughsubstituting otherstones
or throughstealingstones in his temporarypossession,he would lose
family,religious,and communityties. The strengthof theseties makes
possibletransactionsin which trustworthiness is taken forgrantedand
trade can occur with ease. In the absence of these ties, elaborateand
expensivebondingand insurancedeviceswould be necessary-or else the
transactionscould not take place.
2. The InternationalHerald Tribuneof June21-22, 1986, contained
an article on page 1 about South Korean studentradical activists.It
describesthe developmentof such activism:"Radical thoughtis passed
on in clandestine'studycircles,'groupsof studentswho may come from
the same high school or hometownor church.These studycircles. . .
serveas the basic organizationalunitfordemonstrations and otherpro-
tests.To avoid detection,membersof different groupsnevermeet,but
communicatethroughan appointedrepresentative."
This descriptionof the basis of organizationof thisactivismillustrates
social capital of two kinds. The "same high school or hometownor
church"providessocial relationson which the "studycircles"are later
built.The studycirclesthemselvesconstitutea formof social capital-a
cellularformoforganizationthatappearsespeciallyvaluable forfacilitat-
ing oppositionin any politicalsystemintolerantof dissent.Even where
politicaldissentis tolerated,certainactivitiesare not,whethertheactivi-
tiesare politicallymotivatedterrorism or simplecrime.The organization
thatmakes possibletheseactivitiesis an especiallypotentformof social
capital.
3. A motherof six children,who recentlymoved with husband and
childrenfromsuburbanDetroitto Jerusalem,describedas one reasonfor
doing so the greaterfreedomher youngchildrenhad in Jerusalem.She
feltsafe in lettingher eightyear old take the six yearold acrosstownto
schoolon the citybus and feltherchildrento be safein playingwithout
supervisionin a citypark, neitherof whichshe feltable to do whereshe
lived before.
The reasonforthisdifference can be describedas a differencein social
capital available in Jerusalemand suburbanDetroit.In Jerusalem,the

S99

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

normativestructureensures that unattendedchildrenwill be "looked

after"by adults in the vicinity,whileno such normativestructure exists
in mostmetropolitan areas oftheUnitedStates.One can say thatfamilies
have available to themin Jerusalemsocial capital thatdoes not existin
metropolitan areas of the UnitedStates.
4. In the Kahn El Khalili marketof Cairo, the boundariesbetween
merchantsare difficult foran outsiderto discover.The ownerof a shop
thatspecializesin leatherwill, whenqueriedabout whereone can finda
certainkindofjewelry,turnout to sell thatas well-or, whatappearsto
be nearlythe same thing,to have a close associatewho sellsit, to whom
he will immediatelytake the customer.Or he will instantlybecome a
moneychanger,althoughhe is nota moneychanger,merelybyturningto
his colleaguea few shops down. For some activities,such as bringinga
customerto a friend'sstore,thereare commissions;forothers,such as
moneychanging,merelythecreationof obligations.Familyrelationsare
importantin the market,as is the stabilityof proprietorship. The whole
marketis so infusedwithrelationsofthesortI have describedthatit can
be seen as an organization,no less so than a departmentstore.Alterna-
tively,one can see the marketas consistingof a set of individualmer-
chants,each havingan extensivebodyofsocial capitalon whichto draw,
throughthe relationshipsof the market.
The examplesabove have shownthevalue of social capitalfora num-
ber of outcomes,botheconomicand noneconomic.There are, however,
certainpropertiesof social capital thatare importantforunderstanding
how it comesintobeingand how it is employedin thecreationofhuman
capital. First,a comparisonwithhuman capital, and thenan examina-
tionof different formsof social capital, will be helpfulforseeingthese.

HUMAN CAPITAL AND SOCIAL CAPITAL

Probablythe mostimportantand mostoriginaldevelopmentin the eco-
nomicsofeducationin thepast 30 yearshas beentheidea thattheconcept
of physicalcapital as embodiedin tools,machines,and otherproductive
equipmentcan be extendedto includehumancapitalas well (see Schultz
1961; Becker 1964). Just as physicalcapital is created by changes in
materialsto formtoolsthatfacilitateproduction,humancapitalis created
by changesin personsthatbringabout skillsand capabilitiesthatmake
themable to act in new ways.
Social capital, however,comesabout throughchangesin the relations
amongpersonsthatfacilitateaction.If physicalcapitalis whollytangible,
being embodiedin observablematerialform,and humancapital is less
tangible,being embodied in the skills and knowledgeacquired by an
individual,social capital is less tangibleyet,forit existsin the relations

SLOO

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

amongpersons.Justas physicalcapitaland humancapitalfacilitatepro-

ductiveactivity,social capital does as well. For example,a groupwithin
which thereis extensivetrustworthiness and extensivetrustis able to
accomplishmuchmorethan a comparablegroupwithoutthattrustwor-
thinessand trust.

FORMS OF SOCIAL CAPITAL

The value of the concept of social capital lies firstin the fact that it
identifies certainaspectsof social structureby theirfunctions, just as the
concept"chair" identifiescertainphysicalobjects by theirfunction,de-
spite differences in form,appearance, and construction.The function
identified bytheconceptof"social capital"is thevalue oftheseaspectsof
social structureto actorsas resourcesthattheycan use to achieve their
interests.
By identifying this functionof certainaspects of social structure,the
conceptofsocial capitalconstitutes bothan aid in accountingfordifferent
outcomesat thelevel of individualactorsand an aid towardmakingthe
micro-to-macro transitionswithoutelaboratingthe social structuralde-
tails throughwhichthisoccurs. For example,in characterizing theclan-
destinestudy circles of South Korean radical studentsas constituting
social capital thatthesestudentscan use in theirrevolutionary activities,
we assertthatthe groupsconstitutea resourcethataids in movingfrom
individualprotestto organizedrevolt.If, in a theoryofrevolt,a resource
thataccomplishesthistaskis heldto be necessary,thenthesestudycircles
are groupedtogetherwith those organizationalstructures, having very
different origins,thathave fulfilledthesame function forindividualswith
revolutionary goals in othercontexts,such as theComite'sd'action lyce'en
oftheFrenchstudentrevoltof 1968or theworkers'cellsin tsaristRussia
describedand advocated by Lenin ([1902] 1973).
It is true,ofcourse,thatforotherpurposesone wantsto investigate the
detailsof such organizationalresources,to understandthe elementsthat
are criticalto theirusefulnessas resourcesforsuch a purpose,and to
examinehow theycame intobeingin a particularcase. But theconceptof
social capital allows takingsuch resourcesand showingtheway theycan
be combinedwith otherresourcesto producedifferent system-level be-
havioror, in othercases, different outcomesforindividuals.Although,
forthese purposes,social capital constitutesan unanalyzedconcept,it
signalsto the analystand to the readerthatsomethingof value has been
producedforthoseactorswho have thisresourceavailable and thatthe
value depends on social organization.It thenbecomesa second stagein
theanalysisto unpacktheconcept,to discoverwhatcomponentsofsocial
organizationcontributeto the value produced.

s1o0

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

In previouswork, Lin (1988) and De Graf and Flap (1988), froma

perspectiveof methodologicalindividualismsimilarto thatused in this
paper,have shownhow informalsocial resourcesare used instrumentally
in achievingoccupationalmobilityin the UnitedStates and, to a lesser
extent,in WestGermanyand theNetherlands.Lin focusedon social ties,
especially"weak" ties,in thisrole. Here, I want to examinea varietyof
resources,all of whichconstitutesocial capital foractors.
Beforeexaminingempiricallythevalue ofsocial capitalin thecreation
ofhumancapital,I will go moredeeplyintoan examinationofjust what
it is about social relationsthatcan constituteusefulcapital resourcesfor
individuals.

Obligations,Expectations,and Trustworthiness
of Structures
If A does somethingforB and trustsB to reciprocatein the future,this
establishesan expectationin A and an obligationon the partof B. This
obligationcan be conceivedas a creditslip held byA forperformance by
B. If A holdsa largenumberofthesecreditslips,fora numberofpersons
withwhomA has relations,thentheanalogyto financialcapitalis direct.
These creditslips constitutea large body of creditthatA can call in if
necessary-unless,ofcourse,theplacementoftrusthas beenunwise,and
theseare bad debtsthatwill not be repaid.
In somesocial structures, it is said that"peopleare alwaysdoingthings
foreach other."There are a large numberof thesecreditslips outstand-
ing,oftenon bothsidesofa relation(forthesecreditslipsappearoftennot
to be completelyfungibleacrossareas ofactivity,so thatcreditslipsofB
held byA and thoseofA held byB are notfullyused to canceleach other
out). The El Khalili marketin Cairo, describedearlier,constitutesan
extremecase of such a social structure.In othersocial structures where
individualsare moreself-sufficient and depend on each otherless, there
are fewerof thesecreditslips outstandingat any time.
This formofsocial capitaldependson two elements:trustworthiness of
thesocial environment, whichmeansthatobligationswillbe repaid,and
theactual extentofobligationsheld. Social structures differin boththese
dimensions,and actorswithinthe same structure differin the second.A
case thatillustratesthevalue ofthetrustworthiness oftheenvironment is
thatof the rotating-creditassociationsof SoutheastAsia and elsewhere.
These associationsare groupsoffriendsand neighborswho typically meet
monthly,each personcontributing to a centralfundthatis thengivento
one of the members(throughbiddingor by lot), until,aftera numberof
months,each ofthen personshas made n contributions and receivedone
payout.As Geertz(1962) pointsout, theseassociationsserveas efficient

S102

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

institutions foramassing savingsforsmall capital expenditures,an im-

portantaid to economicdevelopment.
But withouta highdegreeof trustworthiness among the membersof
the group,the institutioncould not exist-for a personwho receivesa
payout early in the sequence of meetingscould abscond and leave the
otherswitha loss. For example,one could not imaginea rotating-credit
associationoperatingsuccessfully in urbanareas markedbya highdegree
of social disorganization-or,in otherwords,by a lack of social capital.
Differencesin social structuresin both dimensionsmay arise for a
varietyof reasons.There are differences in the actual needs thatpersons
have forhelp,in theexistenceofothersourcesofaid (suchas government
welfareservices),in the degreeof affluence(which reducesaid needed
fromothers),in culturaldifferences in the tendencyto lend aid and ask
foraid (see Banfield1967)in theclosureofsocialnetworks,in thelogistics
of social contacts(see Festinger,Schachter,and Back 1963), and other
factors.Whateverthe source, however,individualsin social structures
withhighlevels of obligationsoutstandingat any timehave moresocial
capital on whichtheycan draw. The densityof outstandingobligations
means, in effect,that the overall usefulnessof the tangibleresourcesof
that social structureis amplifiedby theiravailabilityto otherswhen
needed.
Individualactorsin a social systemalso differin thenumberof credit
slipsoutstandingon whichtheycan draw at anytime.The mostextreme
examples are in hierarchicallystructuredextendedfamilysettings,in
which a patriarch(or "godfather")holds an extraordinarily large set of
obligationsthat he can call in at any timeto get what he wants done.
Near this extremeare villages in traditionalsettingsthat are highly
stratified, with certainwealthyfamilieswho, because of theirwealth,
have builtup extensivecreditsthattheycan call in at any time.
Similarly,in politicalsettingssuch as a legislature,a legislatorin a
positionwithextraresources(such as theSpeakeroftheHouse ofRepre-
sentativesor theMajorityLeader oftheSenatein theU.S. Congress)can,
by effectiveuse of resources,build up a set of obligationsfromother
legislatorsthat makes it possible to get legislationpassed that would
otherwisebe stymied.This concentration ofobligationsconstitutes social
capitalthatis usefulnotonlyforthispowerfullegislatorbutusefulalso in
gettingan increasedlevel of action on the part of a legislature.Thus,
those membersof legislaturesamong whom such creditsare extensive
shouldbe morepowerfulthanthosewithoutextensivecreditsand debits
because theycan use thecreditsto producebloc votingon manyissues.It
is well recognized,forexample,thatin theU.S. Senate,somesenatorsare
membersof what is called "the Senate Club," whileothersare not. This

S103

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

in effectmeansthatsome senatorsare embeddedin thesystemof credits

and debits, while others,outside the "Club," are not. It is also well
recognizedthatthosein theClub are morepowerfulthanthoseoutsideit.

InformationChannels
An importantformof social capital is the potentialforinformation that
inheresin social relations.Information is importantin providinga basis
foraction. But acquisitionof information is costly.At a minimum,it
requiresattention,whichis alwaysin scarcesupply.One meansbywhich
information can be acquired is by use of social relationsthat are main-
tained forotherpurposes. Katz and Lazarsfeld(1955) showed how this
operatedforwomenin severalareas of lifein a midwesterncityaround
1950.Theyshowedthata womanwithan interestin beingin fashion,but
no interestin beingon theleadingedge of fashion,used friendswho she
knewkeptup withfashionas sourcesofinformation. Similarly,a person
who is not greatlyinterestedin currenteventsbut who is interestedin
beinginformedabout importantdevelopmentscan save thetimeofread-
ing a newspaperby dependingon spouse or friendswho pay attentionto
such matters.A social scientistwho is interestedin beingup-to-dateon
researchin related fieldscan make use of everydayinteractionswith
colleaguesto do so, but onlyin a universityin whichmostcolleagueskeep
up-to-date.
All theseare examplesofsocial relationsthatconstitute a formofsocial
capital thatprovidesinformation thatfacilitatesaction. The relationsin
thiscase are notvaluable forthe"creditslips"theyprovidein theformof
obligationsthat one holds forothers'performances or forthe trustwor-
thinessof the otherpartybut merelyforthe information theyprovide.

Normsand EffectiveSanctions
When a normexistsand is effective,it constitutesa powerful,though
sometimesfragile,formof social capital. Effectivenormsthat inhibit
crimemake it possibleto walk freelyoutsideat nightin a cityand enable
old personsto leave theirhouseswithoutfearfortheirsafety.Normsin a
community thatsupportand provideeffective rewardsforhighachieve-
mentin school greatlyfacilitatethe school'stask.
A prescriptive normwithina collectivity thatconstitutesan especially
importantformof social capital is the normthatone shouldforgoself-
interestand act in the interestsof the collectivity.A normof this sort,
reinforcedby social support,status, honor,and otherrewards,is the
social capital thatbuildsyoungnations(and thendissipatesas theygrow
older),strengthens familiesby leadingfamilymembersto act selflesslyin

S104

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

"thefamily's"interest,facilitatesthedevelopmentofnascentsocialmove-
mentsthrougha small groupofdedicated,inward-looking, and mutually
rewardingmembers,and in generalleads personsto workforthepublic
good. In some of thesecases, the normsare internalized;in others,they
are largelysupportedthroughexternalrewardsforselflessactions and
disapproval for selfishactions. But, whethersupportedby internalor
externalsanctions,normsof this sortare importantin overcomingthe
public goods problemthatexistsin collectivities.
As all theseexamplessuggest,effective normscan constitutea power-
ful formof social capital. This social capital, however,like the forms
describedearlier,notonlyfacilitatescertainactions;it constrainsothers.
A communitywithstrongand effective normsabout youngpersons'be-
havior can keep themfrom"having a good time." Normsthat make it
possibleto walk alone at nightalso constrainthe activitiesof criminals
(and in some cases ofnoncriminals normsthat
as well). Even prescriptive
rewardcertainactions,like thenormin a community thatsaysthata boy
who is a good athleteshould go out forfootball,are in effectdirecting
energyaway fromotheractivities.Effectivenormsin an area can reduce
innovativenessin an area, notonlydeviantactionsthatharmothersbut
also deviant actions that can benefiteveryone. (See Merton [1968,
pp. 195-203] fora discussionof how thiscan come about.)

SOCIAL STRUCTURE THAT FACILITATES SOCIAL CAPITAL

All social relationsand social structuresfacilitatesome formsof social
capital; actorsestablishrelationspurposefully and continuethemwhen
theycontinueto providebenefits.Certainkindsof social structure, how-
ever,are especiallyimportantin facilitatingsomeformsof social capital.

Closureof Social Networks

One propertyof social relationson whicheffective normsdependis what
I will call closure. In general, one can say that a necessarybut not
sufficientconditionforthe emergenceof effectivenormsis action that
imposesexternaleffectson others(see Ullmann-Margalit1977; Coleman
1987). Norms arise as attemptsto limitnegativeexternaleffectsor en-
couragepositiveones. But, in manysocial structures wherethesecondi-
tionsexist,normsdo not come intoexistence.The reasonis what can be
describedas lack of closureof the social structure.Figure 1 illustrates
why.In an open structure like thatoffigurela, actorA, havingrelations
withactorsB and C, can carryout actionsthatimposenegativeexter-
nalitieson B or C or both. Since theyhave no relationswithone another,
but with othersinstead(D and E), thentheycannotcombineforcesto

S105

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

D E

B C B C

A A

(a) (b)
FIG. 1.-Networkwithout
(a) andwith(b) closure

sanctionA in orderto constraintheactions.Unless eitherB or C alone is

sufficientlyharmed and sufficiently powerfulvis-a-visA to sanction
alone,A's actionscan continueunabated.In a structure withclosure,like
thatoffigurelb, B and C can combineto providea collectivesanction,or
eithercan reward the otherfor sanctioningA. (See Merry[1984] for
examplesoftheway gossip,whichdependson closureofthesocial struc-
ture,is used as a collectivesanction.)
In the case of normsimposedby parentson children,closureof the
structurerequiresa slightlymore complexstructure,which I will call
intergenerational closure.Intergenerational closuremay be describedby
a simplediagramthatrepresents relationsbetweenparentand childand
relationsoutsidethe family.Considerthe structure of two communities,
represented by figure2. The verticallinesrepresentrelationsacrossgen-
erations,betweenparentand child,while the horizontallines represent
relationswithina generation.The pointlabeled A in bothfigure2a and
figure2b representstheparentof childB, and thepointlabeled D repre-
sents the parentof child C. The lines betweenB and C representthe
relationsamongchildrenthatexistwithinany school.Althoughtheother
relationsamong childrenwithinthe school are not shown here, there
existsa high degreeof closureamong peers, who see each otherdaily,
have expectationstoward each other,and develop normsabout each
other'sbehavior.
The two communitiesdiffer,however,in the presenceor absence of
linksamong the parentsof childrenin the school. For the school repre-
sentedbyfigure2b, thereis intergenerational closure;forthatrepresented
by figure2a, thereis not. To put it colloquially,in thelowercommunity
represented by 2b, the parents'friendsare the parentsof theirchildren's
friends.In the other,theyare not.

S106

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

E A D E

B C

(a)

A D

B C
(b)

FIG.2.-Networkinvolving parents (B, C) without

(A,D) andchildren (a) and
with(b) intergenerational
closure.

The consequence of this closure is, as in the case of the wholesale

diamondmarketor in othersimilarcommunities, a set of effectivesanc-
tionsthatcan monitorand guidebehavior.In thecommunity in figure2b,
parentsA and D can discusstheirchildren'sactivitiesand come to some
consensusabout standardsand about sanctions.ParentA is reinforced by
parentD in sanctioninghis child'sactions;beyondthat,parentD consti-
tutesa monitornotonlyforhis own child,C, but also fortheotherchild,
B. Thus, theexistenceofintergenerational closureprovidesa quantityof
social capitalavailable to each parentin raisinghis children-not onlyin
mattersrelatedto school but in othermattersas well.
Closureofthesocial structure is importantnotonlyfortheexistenceof
effective normsbut also foranotherformof social capital: the trustwor-
thinessof social structures thatallows theproliferationofobligationsand
expectations.Defectionfroman obligationis a formofimposinga nega-
tive externalityor another.Yet, in a structure withoutclosure,it can be
effectivelysanctioned,ifat all, onlybythepersonto whomtheobligation
is owed. Reputationcannot arise in an open structure,and collective

S107

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

sanctionsthatwould ensuretrustworthiness
cannotbe applied. Thus, we
in a social structure.
may say thatclosurecreatestrustworthiness

AppropriableSocial Organization
Voluntaryorganizationsare broughtinto being to aid some purposeof
thosewho initiatethem.In a housingprojectbuiltduringWorldWar II
in an easterncityof the UnitedStates, therewere manyphysicalprob-
lemscaused bypoorconstruction: faultyplumbing,crumblingsidewalks,
and otherdefects(Merton,n.d.). Residentsorganizedto confrontthe
buildersand to address theseproblemsin otherways. Later, when the
problemsweresolved,theorganizationremainedas available social capi-
tal thatimprovedthequalityoflifeforresidents.Residentshad resources
available thattheyhad seen as unavailablewheretheyhad lived before.
(For example,despitethe factthatthe numberof teenagersin the com-
munitywas smaller,residentswere morelikelyto expresssatisfaction
withthe availabilityof teenagebabysitters.)
Printersin the New York TypographicalUnion who were monotype
operatorsformeda MonotypeClub as a social club (Lipset, Trow, and
Coleman 1956). Later, as employerslooked formonotypeoperatorsand
as monotypeoperatorslooked forjobs, bothfoundthisorganizationan
effective employmentreferralserviceand appropriatedthe organization
forthispurpose.Stilllater,whenthe ProgressivePartycame intopower
in theNew York Union, theMonotypeClub servedas an organizational
resourceforthe IndependentPartyas it leftoffice.The MonotypeClub
subsequentlyservedas an importantsourceofsocial capitalfortheInde-
pendentsto sustainthepartyas an organizedoppositionwhileit was out
of office.
In theexampleofSouthKorean studentradicalsused earlier,thestudy
circleswere describedas consistingof groupsof studentsfromthe same
highschoolor hometownor church.Here, as in theearlierexamples,an
organizationthatwas initiatedforone purposeis available forappropria-
tionforotherpurposes,constituting importantsocial capitalfortheindi-
vidual members,who have available to themtheorganizationalresources
necessaryforeffectiveopposition.These examplesillustratethe general
point,thatorganization,once broughtinto existenceforone set of pur-
poses, can also aid others,thus constituting social capital available for
use.
It is possibleto gain insightintosomeofthewaysin whichclosureand
appropriablesocial organizationprovidesocial capitalbyuse ofa distinc-
tionmade byMax Gluckman(1967) betweensimplexand multiplexrela-

S108

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

tions.3In the latter,personsare linkedin morethanone context(neigh-

bor,fellowworker,fellowparent,coreligionist, etc.),whilein theformer,
personsare linkedthroughonlyone of theserelations.The centralprop-
ertyof a multiplexrelationis thatit allows the resourcesofone relation-
ship to be appropriatedfor use in others.Sometimes,the resourceis
merelyinformation, as whentwoparentswho see each otheras neighbors
exchangeinformation about theirteenagers'activities;sometimes,itis the
obligationsthat one personowes a second in relationshipX, whichthe
secondpersoncan use to constrainthe actionsof thefirstin relationship
Y. Often,itis resourcesin theformofotherpersonswho have obligations
in one contextthat can be called on to aid when one has problemsin
anothercontext.

SOCIAL CAPITAL IN THE CREATION OF HUMAN CAPITAL

The precedingpages have been directedtowarddefiningand illustrating
social capital in general.But thereis one effectof social capital that is
especiallyimportant:its effecton the creationof human capital in the
nextgeneration.Both social capitalin thefamilyand social capitalin the
community play rolesin thecreationofhumancapitalin therisinggener-
ation. I will examineeach of thesein turn.

Social Capital in the Family

Ordinarily,in theexaminationoftheeffects ofvariousfactorson achieve-
mentin school,"familybackground"is considereda singleentity,distin-
guishedfromschoolingin its effects.But thereis not merelya single
"familybackground";familybackgroundis analyticallyseparableintoat
least threedifferent components:financialcapital, human capital, and
social capital. Financial capital is approximately
measuredby the fami-
ly's wealth or income. It providesthe physicalresourcesthat can aid
achievement:a fixedplace in the home for studying,materialsto aid
learning,the financialresourcesthat smoothfamilyproblems.Human
capitalis approximately measuredbyparents'educationand providesthe
potentialfor a cognitiveenvironmentfor the child that aids learning.
Social capital withinthe familyis different fromeitherof these. Two
exampleswill give a sense of what it is and how it operates.
JohnStuartMill, at an age beforemost childrenattendschool, was
taughtLatin and Greekby his father,JamesMill, and laterin childhood

3 I am especially meofGluckman's
toSusanShapiroforreminding
grateful distinction
and pointingout therelevanceofit to myanalysis.

S109

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

would discusscriticallywithhis fatherand withJeremyBenthamdrafts

of his father'smanuscripts.JohnStuartMill probablyhad no extraordi-
narygeneticendowments,and his father'slearningwas no moreexten-
sive thanthatof some othermen of thetime.The centraldifference was
the time and effortspent by the fatherwith the child on intellectual
matters.
In one publicschooldistrictin theUnitedStateswheretextsforschool
use were purchased by children'sfamilies, school authoritieswere
puzzledto discoverthata numberofAsian immigrant familiespurchased
two copies of each textbookneeded by the child. Investigationrevealed
thatthefamilypurchasedthesecondcopyforthemotherto studyin order
to help her child do well in school. Here is a case in whichthe human
capital of the parents, at least as measured traditionallyby years of
schooling,is low, but the social capital in the familyavailable forthe
child'seducationis extremely high.
These examplesillustratethe importanceof social capital withinthe
familyfor a child's intellectualdevelopment.It is of course true that
childrenare stronglyaffectedby the human capital possessed by their
parents.But thishuman capital may be irrelevantto outcomesforchil-
drenifparentsare not an importantpartoftheirchildren'slives,iftheir
human capital is employedexclusivelyat workor elsewhereoutsidethe
home. The social capital of the familyis the relationsbetweenchildren
and parents(and, when familiesinclude othermembers,relationships
withthemas well). That is, ifthe humancapital possessedby parentsis
not complementedby social capital embodied in familyrelations,it is
irrelevantto the child's educationalgrowththat the parenthas a great
deal, or a small amount,of humancapital.4
I will notdifferentiate
hereamongtheformsofsocial capitaldiscussed
earlier,but will attemptmerelyto measurethe strengthof the relations
betweenparentsand childas a measureof the social capital available to
thechildfromtheparent.Nor will I use theconceptin thecontextofthe
paradigmof rationalaction,as, forexample,is oftendone in use of the
conceptof humancapital to examinetheinvestments in educationthata
rationalperson would make. A portionof the reason forthis lies in a
propertyof much social capital not shown by mostformsof capital (to

IThe complementarity ofhumancapitaland socialcapitalin thefamilyfora child's

development suggeststhatthe statisticalanalysisthatexaminestheeffects
of these
quantities
shouldtakea particular form.Thereshouldbe an interaction
termbetween
humancapital(parents'education)and socialcapital(somecombination ofmeasures
such as two parentsin the home,numberof siblings,and parents'expectations for
child'seducation).In theanalysisreported, here,however,a simpleadditivemodel
withoutinteraction was used.

SilO

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

which I will turnin a later section):its public goods character,which

leads to underinvestment.
Social capitalwithinthefamilythatgivesthechildaccess to theadult's
human capital depends both on the physicalpresenceof adults in the
familyand on theattentiongivenbytheadultsto thechild.The physical
absence of adults may be describedas a structuraldeficiency in family
social capital. The most prominentelementof structuraldeficiencyin
modernfamiliesis thesingle-parent family.However,thenuclearfamily
itself,in whichone or bothparentsworkoutsidethehome,can be seenas
structurally deficient,lackingthesocial capitalthatcomeswiththepres-
ence ofparentsduringtheday, or withgrandparents or auntsand uncles
in or near the household.
Even ifadultsare physicallypresent,thereis a lack ofsocial capitalin
thefamilyifthereare not strongrelationsbetweenchildrenand parents.
The lack ofstrongrelationscan resultfromthechild'sembeddednessin a
youthcommunity, fromthe parents'embeddednessin relationships with
otheradults thatdo not crossgenerations,or fromothersources.What-
ever the source,it means thatwhateverhumancapital existsin thepar-
ents,thechilddoes notprofitfromit becausethesocial capitalis missing.
The effectsofa lack ofsocial capitalwithinthefamilydifferfordiffer-
ent educationaloutcomes.One forwhichit appears to be especiallyim-
portantis droppingout of school. With the High School and Beyond
sample of studentsin high schools,table 1 shows the expecteddropout
ratesforstudentsin different typesoffamilieswhenvariousmeasuresof
social and humancapital in thefamilyand a measureof social capitalin
the communityare controlledstatistically.5 An explanationis necessary
forthe use of numberof siblingsas a measureof lack of social capital.
The numberof siblingsrepresents,in this interpretation, a dilutionof
adult attentionto the child. This is consistentwith researchresultsfor

5 The analysisis carriedoutbyuse ofa weighted logisticmodelwitha randomsample

of4,000students fromthepublicschoolsin thesample.The variablesincludedin the
modelas measuresofthefamily's financial,
human,and socialcapitalweresocioeco-
nomicstatus(a singlevariableconstructed of parents'education,parents'income,
father'soccupationalstatus,and householdpossessions),race, Hispanicethnicity,
numberofsiblings,numberofchangesin schooldue tofamilyresidential movessince
fifthgrade,whethermotherworkedbeforethechildwas in school,mother's expecta-
tionof child'seducationalattainment, frequency of discussionswithparentsabout
personalmatters,and presenceof both parentsin the household.The regression
coefficients
and asymptotic standarderrorsaregivenintheApp. tableAl. Ananalysis
withmoreextensivestatisticalcontrols,includingsuch thingsas gradesin school,
homework, and numberof absences,is reportedin Hoffer(1986,table25), but the
effectsreportedin table1 and subsequenttextare essentiallyunchangedexceptfora
reducedeffectof mother'sexpectations. The resultsreported hereand subsequently
are takenfromHoffer(1986)and fromColemanand Hoffer(1987).

Sill

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

TABLE 1

DROPOUT RATES BETWEEN SPRING, GRADE 10, AND SPRING, GRADE 12, FOR STUDENTS
WHOSE FAMILIES DIFFER IN SOCIAL CAPITAL, CONTROLLING FOR HUMAN CAPITAL
AND FINANCIAL CAPITAL IN THE FAMILYa

in
Percentage Difference
Dropping Percentage
Out Points

1. Parents'presence:
Two parents ..................... 13.1 6.0
Singleparent ............... .. .................. 19.1
2. Additionalchildren:
One sibling .......... ...... .. ................... 10.8 6.4
Four siblings ......... ...... .. .................. 17.2
3. Parentsand children:
Two parents,one sibling .......... .. ............. 10.1 12.5
One parent,foursiblings .......... .. ............. 22.6 J
4. Mother'sexpectationforchild'seducation:
Expectationof college ............................ 11.6 l 8.6
No expectationof college .......... .. ............. 20.2
5. Three factorstogether:
Two parents,one sibling,motherexpectscollege .... 8.1 22.5
One parent,foursiblings,no collegeexpectation ..... 30.6 J
a takenfrom
Estimates logistic morefullyinApp.tableAl.
reported
regression

measuresof achievementand IQ, which show that test scores decline

with sib position,even when total familysize is controlled,and that
scores decline with numberof childrenin the family.Both resultsare
consistentwiththe view thatyoungersibs and childrenin largefamilies
have less adult attention,whichproducesweakereducationaloutcomes.
Item 1 of table 1 shows that, when otherfamilyresourcesare con-
trolled,the percentageof studentswho drop out betweenspringof the
sophomoreyearand springofthesenioryearis 6 percentagepointshigher
forchildrenfromsingle-parent families.Item 2 of table 1 showsthatthe
rate is 6.4 percentagepoints higherforsophomoreswith foursiblings
than forthose with otherwiseequivalentfamilyresourcesbut onlyone
sibling.Or, takingthesetwo together,we can thinkoftheratioofadults
to childrenas a measureofthesocial capitalin thefamilyavailable forthe
educationof any one of them.Item 3 of table 1 shows thatfora sopho-
morewithfoursiblingsand one parent,and an otherwiseaverage back-
ground,the rate is 22.6%; withone siblingand two parents,the rate is
10.1%-a difference of 12.5 percentagepoints.
Anotherindicatorofadult attentionin thefamily,althoughnot a pure
measureof social capital, is the mother'sexpectationof the child'sgoing
to college. Item 4 of the table shows that,forsophomoreswithoutthis

S1 12

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

parentalexpectation,the rate is 8.6 percentagepointshigherthan for

those with it. With the threesourcesof familysocial capital taken to-
gether,item5 of the table shows thatsophomoreswithone sibling,two
parents,and a mother'sexpectationforcollege(stillcontrolling on other
resourcesof family)have an 8.1% dropoutrate; withfoursiblings,one
parent,and no expectationof the motherforcollege,the rateis 30.6%.
These resultsprovidea less satisfactory testthan if the researchhad
been explicitlydesignedto examine effectsof social capital withinthe
family.In addition,tableAl in theAppendixshowsthatanothervariable
thatshouldmeasuresocial capital in thefamily,thefrequencyoftalking
withparentsabout personalexperiences,showsessentiallyno relationto
droppingout. Nevertheless,takenall together,thedata do indicatethat
social capital in the familyis a resourcefor educationof the family's
children,just as is financialand humancapital.

Social Capital outsidethe Family

The social capital thathas value fora youngperson'sdevelopmentdoes
not residesolelywithinthefamily.It can be foundoutsideas well in the
community consistingofthesocial relationships thatexistamongparents,
in the closureexhibitedby thisstructure of relations,and in theparents'
relationswiththe institutions of the community.
The effectof thissocial capital outsidethe familyon educationalout-
comes can be seen by examiningoutcomesforchildrenwhose parents
differin theparticularsourceofsocial capitaldiscussedearlier,intergen-
erationalclosure.There is not a directmeasureof intergenerational clo-
surein thedata, but thereis a proximateindicator.This is thenumberof
times the child has changed schools because the familymoved. For
familiesthathave moved often,the social relationsthatconstitutesocial
capital are brokenat each move. Whateverthe degreeof intergenera-
tionalclosureavailable to othersin the community, it is not available to
parentsin mobilefamilies.
The logisticregressioncarried out earlierand reportedin table Al
showsthatthecoefficient fornumberofmovessincegrade5 is 10 timesits
standarderror,the variablewiththe strongest overalleffectof any vari-
able in the equation, includingthe measuresof human and financial
capital in the family(socioeconomicstatus)and the crude measuresof
familysocial capital introducedin the earlieranalysis.Translatingthis
into an effecton droppingout gives 11.8% as the dropoutrate if the
familyhas not moved, 16.7% if it has moved once, and 23.1% if it has
moved twice.
In theHigh School and Beyonddata set,anothervariationamongthe
schoolsconstitutes a usefulindicatorof social capital. This is thedistinc-

S113

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

tionsamong public highschools,religiouslybased privatehighschools,

and nonreligiously based privatehighschools.It is the religiouslybased
highschoolsthatare surroundedby a community based on the religious
organization.These familieshave intergenerational closurethatis based
on a multiplexrelation:whateverotherrelationstheyhave, theadultsare
membersof the same religiousbodyand parentsof childrenin the same
school.In contrast,it is theindependentprivateschoolsthatare typically
least surroundedby a community, fortheirstudentbodiesare collections
of students,most of whose familieshave no contact.6The choice of
privateschool formost of these parentsis an individualisticone, and,
althoughtheyback theirchildrenwith extensivehuman capital, they
send theirchildrento theseschoolsdenudedof social capital.
In theHigh School and Beyonddata set,thereare 893 publicschools,
84 Catholic schools, and 27 other private schools. Most of the other
privateschoolsare independentschools,thougha minority have religious
foundations.In thisanalysis,I will at the outsetregardtheotherprivate
schools as independentprivateschools to examinethe effectsof social
capital outsidethe family.
The resultsof these comparisonsare shown in table 2. Item 1 of the
table shows thatthe dropoutratesbetweensophomoreand senioryears
are 14.4% in publicschools,3.4% in Catholicschools,and 11.9% in other
privateschools.What is moststriking is thelow dropoutratein Catholic
schools. The rate is a fourthof thatin the public schoolsand a thirdof
thatin the otherprivateschools.
Adjustingthe dropoutrates fordifferences in student-body financial,
human, and social capital among the three sets of schools by stan-
dardizingthepopulationoftheCatholicschoolsand otherprivateschools
to the student-bodybackgroundsof the public schools shows that the
differences are affectedonlyslightly.Furthermore, thedifferences are not
due to thereligionofthestudentsor to thedegreeofreligiousobservance.
Catholicstudentsin public schoolare onlyslightlyless likelyto dropout
thannon-Catholics.Frequencyofattendanceat religiousservices,which
is itselfa measureof social capital throughintergenerational closure,is
stronglyrelatedto dropoutrate, with 19.5% of public school students
who rarelyor never attenddroppingout comparedwith 9.1% of those
who attendoften.But thiseffectexistsapartfrom,and in additionto, the
effectoftheschool'sreligiousaffiliation.ComparablefiguresforCatholic

6 Data fromthisstudyhave no directmeasuresof the degreeof intergenerational

closureamongtheparentsoftheschoolto supportthisstatement. However,theone
measureofintergenerationalclosurethatdoesexistin thedata,thenumberofresiden-
tialmovesrequiring schoolchangesincegrade5, is consistent
withthestatement.
The
averagenumberof movesforpublicschoolstudentsis .57; forCatholicschoolstu-
dents,.35; and forstudentsin otherprivateschools,.88.

S114

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

TABLE 2
DROPOUT RATES BETWEEN SPRING, GRADE 10, AND SPRING, GRADE 12,
FOR STUDENTS FROM SCHOOLS WITH DIFFERING AMOUNTS OF SOCIAL
CAPITAL IN THE SURROUNDING COMMUNITY

Other
Private
Public Catholic Schools

1. Raw dropout rates ................................ 14.4 3.4 11.9

2. Dropout rates standardized to average public
school sophomorea ............................... 14.4 5.2 11.6

Non-Catholic
Religious Independent

3. Raw dropout rates for studentsb from inde-

pendent and non-Catholic religious
private schools ............... ................... 3.7 10.0

a
The standardizationis based on separatelogisticregressions
forthesetwo setsof schools,usingthe
same variableslistedin n. 5. Coefficients
and meansforthestandardization are in Hoffer(1986, tables5
and 24).
b This tabulationis based on unweighteddata, which is responsibleforthe factthat bothrates are
lowerthan the rate forotherprivateschoolsin item 1 of the table, whichis based on weighteddata.

school studentsare 5.9% and 2.6%, respectively(Coleman and Hoffer

1987, p. 138).
The low dropout rates of the Catholic schools, the absence of low
dropoutratesin the otherprivateschools,and the independenteffectof
frequencyof religiousattendanceall provideevidenceof theimportance
of social capital outsidethe school,in the adult communitysurrounding
it, forthisoutcomeof education.
A further testis possible,fortherewereeightschoolsin the sampleof
non-Catholicprivateschools("otherprivate"in the analysisabove) that
have religiousfoundationsand over 50% of the studentbody of that
religion.Three were Baptist schools,two were Jewish,and threefrom
threeotherdenominations.If the inferenceis correctabout the religious
community's providingintergenerational closureand thus social capital
and about the importanceof social capital in depressingthe chance of
droppingout of high school, these schools also should show a lower
dropoutrate than the independentprivate schools. Item 3 of table 2
showsthattheirdropoutrateis lower,3.7%, essentiallythesame as that
of the Catholic schools.7
I It is also true,thoughnotpresented
here,thatthelack ofsocialcapitalin thefamily
makeslittledifference in dropoutratesin Catholicschools-or, in thetermsI have
used,socialcapitalin thecommunity compensatesinpartforitsabsencein thefamily.
See Colemanand Hoffer(1987,chap. 5).

S115

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

The data presentedabove indicatetheimportanceof social capitalfor

the educationof youth,or, as it mightbe put, the importanceof social
capital in the creationof human capital. Yet thereis a fundamental
differencebetweensocial capitaland mostotherformsofcapitalthathas
strongimplicationsforthe developmentof youth.It is thisdifferenceto
whichI will turnin the nextsection.

PUBLIC GOODS ASPECTS OF SOCIAL CAPITAL

Physicalcapitalis ordinarilya privategood, and propertyrightsmake it
possible for the person who investsin physicalcapital to capturethe
itproduces.Thus, theincentiveto investin physicalcapitalis not
benefits
depressed;thereis not a suboptimalinvestmentin physicalcapital be-
cause those who invest in it are able to capture the benefitsof their
investments.For humancapital also-at least humancapital of thesort
that is produced in schools-the personwho investsthe time and re-
sources in buildingup this capital reaps its benefitsin the formof a
higher-paying job, more satisfyingor higher-status work, or even the
pleasureofgreaterunderstanding ofthesurroundingworld-in short,all
the benefitsthatschoolingbringsto a person.
But most formsof social capital are not like this. For example,the
kindsof social structures thatmake possiblesocial normsand the sanc-
tionsthat enforcethemdo not benefitprimarilythe personor persons
whose effortswould be necessaryto bringthemabout, but benefitall
those who are part of such a structure.For example, in some schools
wherethereexistsa dense set of associationsamongsome parents,these
are the resultof a small numberof persons,ordinarilymotherswho do
not hold full-time jobs outsidethe home. Yet thesemothersthemselves
experienceonlya subsetof the benefitsofthissocial capital surrounding
the school. If one of themdecides to abandon these activities-forex-
ample, to take a full-timejob-this maybe an entirelyreasonableaction
froma personalpointof view and even fromthe pointof view of that
householdwith its children.The benefitsof the new activitymay far
outweighthe losses thatarise fromthedeclinein associationswithother
parentswhose childrenare in the school. But the withdrawalof these
activitiesconstitutesa loss to all thoseotherparentswhose associations
and contactswere dependenton them.
Similarly,the decisionto move froma communityso thatthe father,
forexample,can take a betterjob may be entirelycorrectfromthepoint
of view of that family.But, because social capital consistsof relations
among persons, otherpersons may experienceextensivelosses by the
severanceofthoserelations,a severanceoverwhichtheyhad no control.

S116

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

A partofthoselossesis theweakeningofnormsand sanctionsthataid the

school in its task. For each family,the total cost it experiencesas a
consequenceofthedecisionsit and otherfamiliesmakemayoutweighthe
benefitsof those few decisionsit has controlover. Yet the beneficial
consequencesto thefamilyofthosedecisionsmade bythefamilymayfar
outweighthe minorlosses it experiencesfromthemalone.
It is notmerelyvoluntaryassociations,suchas a PTA, in whichunder-
investmentof this sort occurs. When an individualasks a favorfrom
another,thusincurringan obligation,he does so because it bringshima
neededbenefit;he does notconsiderthatit does theothera benefitas well
by addingto a drawingfundofsocial capital available in a timeofneed.
If the firstindividual can satisfyhis need throughself-sufficiency, or
throughaid fromsomeofficialsourcewithoutincurring an obligation,he
will do so-and thus fail to add to the social capital outstandingin the
community.
Similar statementscan be made with respectto trustworthiness as
social capital. An actorchoosingto keeptrustor not(orchoosingwhether
to devoteresourcesto an attemptto keep trust)is doingso on thebasis of
costs and benefitshe himselfwill experience.That his trustworthiness
willfacilitateothers'actionsor thathislack oftrustworthiness willinhibit
others' actions does not enter into his decision. A similar but more
qualifiedstatementcan be made forinformation as a formof social capi-
tal. An individual who serves as a source of information for another
because he is well informedordinarilyacquires thatinformation forhis
own benefit,notfortheotherswho make use ofhim. (This is notalways
true.As Katz and Lazarsfeld[1955] show, "opinionleaders"in an area
acquireinformation in partto maintaintheirpositionas opinionleaders.)
For normsalso, the statementmustbe qualified.Normsare intention-
ally established,indeed as means of reducingexternalities,and their
benefitsare ordinarilycapturedby thosewho are responsibleforestab-
lishingthem.But thecapabilityofestablishingand maintaining effective
normsdependson propertiesofthesocial structure (suchas closure)over
which one actor does not have controlyet are affectedby one actor's
action.These are propertiesthataffectthestructure's capacityto sustain
effective norms,yetpropertiesthatordinarilydo not enterintoan indi-
vidual's decisionthataffectsthem.
Some formsofsocial capitalhave theproperty thattheirbenefits can be
capturedby thosewho investin them;consequently,rationalactorswill
notunderinvestin thistypeof social capital. Organizationsthatproduce
a privategood constitute theoutstanding example.The resultis thatthere
willbe in societyan imbalancein therelativeinvestment in organizations
thatproduceprivategoods fora marketand thoseassociationsand rela-

S117

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

tionshipsin which the benefitsare not captured-an imbalancein the

sensethat,ifthepositiveexternalities createdby thelatterformof social
capital could be internalized,it would come to existin greaterquantity.
The public goods qualityof most social capital means that it is in a
fundamentally positionwithrespectto purposiveactionthanare
different
mostotherformsof capital. It is an importantresourceforindividuals
and may affectgreatlytheirabilityto act and theirperceivedqualityof
life.They have the capabilityof bringingit intobeing.Yet, because the
benefitsof actionsthatbringsocial capital intobeingare largelyexperi-
enced by personsotherthan the actor, it is oftennot in his interestto
bringit into being. The resultis that most formsof social capital are
createdor destroyedas by-products ofotheractivities.This social capital
arisesor disappearswithoutanyone'swillingit intoor out ofbeingand is
thus even less recognizedand taken accountof in social actionthan its
alreadyintangiblecharacterwould warrant.
There are importantimplicationsof thispublic goods aspect of social
capital that play a part in the developmentof childrenand youth.Be-
cause thesocial structuralconditionsthatovercometheproblemsofsup-
plyingthese public goods-that is, strongfamiliesand strongcommu-
nities-are muchless oftenpresentnow thanin thepast, and promiseto
be even less presentin thefuture,we can expectthat,ceterisparibus,we
confront a decliningquantityof humancapitalembodiedin each succes-
sive generation.The obvious solutionappears to be to attemptto find
ways of overcomingtheproblemof supplyof thesepublicgoods,thatis,
social capital employedforthe benefitof childrenand youth.This very
likelymeans thesubstitution of some kindof formalorganizationforthe
voluntaryand spontaneoussocial organizationthathas in the past been
the major sourceof social capital available to theyoung.

CONCLUSION
In thispaper,I have attemptedto introduceintosocial theorya concept,
"social capital," parallelingthe conceptsof financialcapital, physical
capital, and human capital-but embodiedin relationsamong persons.
This is partof a theoreticalstrategythatinvolvesuse of theparadigmof
rationalactionbut withouttheassumptionofatomisticelementsstripped
of social relationships.I have shown the use of this conceptthrough
demonstrating theeffectofsocial capitalin thefamilyand in thecommu-
nityin aiding the formationof human capital. The singlemeasureof
human capital formationused forthis was one that appears especially
responsiveto the supplyof social capital, remainingin highschooluntil
graduationversus droppingout. Both social capital in the familyand
social capital outsideit, in the adult community surrounding the school,

S118

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Social Capital

showed evidence of considerablevalue in reducingthe probabilityof

droppingout of highschool.
In explicatingtheconceptofsocial capital,threeformswereidentified:
obligationsand expectations,which depend on trustworthiness of the
social environment, information-flow capabilityof the social structure,
and normsaccompaniedby sanctions.A propertysharedby mostforms
it fromotherformsof capital is its
of social capital that differentiates
public good aspect: the actoror actorswho generatesocial capital ordi-
narilycaptureonlya smallpartofitsbenefits, a factthatleads to underin-
vestmentin social capital.

APPENDIX
TABLE Al
LOGISTIC REGRESSION COEFFICIENTS AND ASYMPTOTIC STANDARD ERRORS FOR
EFFECTS OF STUDENT BACKGROUND CHARACTERISTICS ON DROPPING OUT OF HIGH
SCHOOL BETWEEN SOPHOMORE AND SENIOR YEARS, 1980-82, PUBLIC SCHOOL SAMPLE

b SE

Intercept ................................................ -2.305 .169

Socioeconomic status .................. ..................... -.460 .077
Black ................................................ -.161 .162
Hispanic ................................................ .104 .138
Number of siblings .................... .................... .180 .028
Mother worked while child was young ...... ................. - .012 .103
Both parents in household ............ ...................... -.415 .112
Mother's expectation for college .......... ................... -.685 .103
Talk with parents . ......................................... .031 .044
Number of moves since grade 5 .......... ................... .407 .040

SOURCE.-Taken fromHoffer(1986).

REFERENCES
Baker,Wayne. 1983. "FloorTradingand CrowdDynamics."Pp. 107-28 in Social
DynamicsofFinancialMarkets,editedby PatriciaAdlerand PeterAdler.Green-
wich,Conn.:JAI.
Banfield,Edward. 1967. The Moral Basis ofa BackwardSociety.New York: Free
Press.
Becker,Gary. 1964.HumanCapital.New York:NationalBureauof EconomicRe-
search.
Ben-Porath, Yoram. 1980."The F-Connection:Families,Friends,and Firmsand the
Organizationof Exchange."Populationand Development Review6:1-30.
Black, R. D. C., A. W. Coats, and C. D. W. Goodwin,eds. 1973. The Marginal
Revolutionin Economics.Durham,N.C.: Duke University Press.
Blau, Peter.1964.Exchangeand Powerin Social Life.New York:Wiley.

S1 19

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AmericanJournalof Sociology

Coleman,JamesS. 1986a."SocialTheory,SocialResearch,and a TheoryofAction."

American JournalofSociology91:1309-35.
. 1986b.IndividualInterestsand CollectiveAction.Cambridge:Cambridge
University Press.
* 1987."Normsas SocialCapital."Pp. 133-55inEconomicImperalism, edited
by GerardRadnitzkyand PeterBernholz.New York: Paragon.
Coleman,J. S., and T. B. Hoffer.1987.Public and PrivateSchools:The Impactof
Communities. New York: Basic.
DeGraaf,Nan Dirk, and HendrikDerk Flap. 1988. "Witha LittleHelp fromMy
Friends."Social FQrces,vol. 67 (in press).
Festinger, Leon, StanleySchachter,and KurtBack. 1963.Social Pressuresin Infor-
mal Groups.Stanford, Calif.:StanfordUniversity Press.
Geertz,Clifford.1962."The RotatingCreditAssociation: A 'MiddleRung'in Devel-
opment."EconomicDevelopment and CulturalChange10:240-63.
Gluckman, Max. 1967.TheJudicialProcessamongtheBarotseofNorthern Rhodesia,
2d ed. Manchester: ManchesterUniversity Press.
Gouldner,Alvin.1960."The NormofReciprocity: A Preliminary Statement."Ameri-
can SociologicalReview25:161-78.
Granovetter, Mark. 1985."EconomicAction,Social Structure, and Embeddedness."
American JournalofSociology91:481-510.
Hoffer, T. B. 1986.EducationalOutcomesin PublicandPrivateHighSchools.Ph.D.
dissertation.University of Chicago,Department of Sociology.
Homans,George.1974.Social Behavior:Its Elementary Forms,rev.ed. New York:
Harcourt,Brace & World.
Katz, E., and P. Lazarsfeld.1955.PersonalInfluence.New York:Free Press.
Lenin,V. I. (1902) 1973. WhatIs To Be Done. Peking:ForeignLanguagePress.
Lin, Nan. 1988."SocialResourcesand SocialMobility: A Structural TheoryofStatus
Attainment." In Social Mobilityand Social Structure, editedby RonaldBreiger.
Cambridge:CambridgeUniversity Press.
Lipset,Seymour,M. Trow, and J. Coleman. 1956. UnionDemocracy.New York:
Free Press.
Merry,Sally,E. 1984. "Rethinking Gossipand Scandal." Pp. 271-302in Towarda
GeneralTheoryofSocial Control.Vol. 1, Fundamentals, editedbyDonald Black.
New York: Academic.
Merton,RobertK. 1968.Social Theoryand Social Structure, 2d ed. New York:Free
Press.
. n.d. "Studyof WorldWar II HousingProjects."Unpublishedmanuscript.
ColumbiaUniversity, Department of Sociology.
Schultz,Theodore.1961. "Investment in Human Capital."AmericanEconomicRe-
view51 (March):1-17.
Ullmann-Margalit, Edna. 1977.TheEmergenceofNorms.Oxford:Clarendon.
Williamson,Oliver.1975.Marketsand Hierarchies.New York: Free Press.
. 1981. "The Economicsof Organization: The TransactionCost Approach."
American JournalofSociology87:548-77.
Wrong,Dennis.1961."The Oversocialized Conception ofMan inModernSociology."
AmericanSociologicalReview26:183-93.

S120

This content downloaded from 141.218.001.105 on July 30, 2016 17:22:51 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).