ECONOMICS AND IDENTITY* R I. I

39
ECONOMICS AND IDENTITY* GEORGE A. AKERLOF AND RACHEL E. KRANTON This paper considers how identity, a person’s sense of self, affects economic outcomes. We incorporate the psychology and sociology of identity into an economic model of behavior. In the utility function we propose, identity is associated with different social categories and how people in these categories should behave. We then construct a simple game-theoretic model showing how identity can affect individual interactions. The paper adapts these models to gender discrimination in the workplace, the economics of poverty and social exclusion, and the household division of labor. In each case, the inclusion of identity substantively changes conclusions of previous economic analysis. I. INTRODUCTION This paper introduces identity—a person’s sense of self—into economic analysis. Identity can account for many phenomena that current economics cannot well explain. It can comfortably resolve, for example, why some women oppose ‘‘women’s rights,’’ as seen in microcosm when Betty Friedan was ostracized by fellow suburban * The authors especially wish to thank Abdeslam Maghraoui for his contin- ued help and insights and Michael Ash, Jennifer Eichberger, and Cyd Fremmer for invaluable research assistance. Henry Aaron, William Dickens, Claudia Goldin, Edward Glaeser, Lawrence Katz, Robert Merton, Anand Swamy, and an anony- mous referee made extensive comments on earlier drafts for which the authors are particularly grateful. They also thank Robert Akerlof, Abhijit Banerjee, Kaushik Basu, Paul Beaudry, Samuel Bowles, Robert Boyd, Gary Burtless, Alessandra Casella, Catherine Eckel, Stuart Elliott, Gary Fields, Pierre Fortin, James Foster, Richard Harris, Victoria Hattam, Peter Howett, Aurora Jackson, Kevin Lang, George Loewenstein, Glenn Loury, Michael Kremer, David Laibson, Janet Pack, Matthew Rabin, Francisco Rodriguez, Paul Romer, Eric Verhoogen, Eric Wanner, Kent Weaver, Robin Wells, Janet Yellen, and Peyton Young for help and comments. George Akerlof is grateful to the Canadian Institute for Advanced Research, the MacArthur Foundation, the Brookings Institution, and the National Science Foundation, under research grant number SBR 97-09250, for nancial support. Rachel Kranton expresses her gratitude to the Russell Sage Foundation where she was a Visiting Scholar for 1997–1998. 2000 by the President and Fellows of Harvard College and the Massachusetts Institute of Technology. The Quarterly Journal of Economics, August 2000 715

Transcript of ECONOMICS AND IDENTITY* R I. I

ECONOMICS AND IDENTITY*

GEORGE A. AKERLOF AND RACHEL E. KRANTON

This paper considers how identity, a person’s sense of self, affects economicoutcomes. We incorporate the psychology and sociology of identity into an economicmodel of behavior. In the utility function we propose, identity is associated withdifferent social categories and how people in these categories should behave. Wethen construct a simple game-theoretic model showing how identity can affectindividual interactions. The paper adapts these models to gender discrimination inthe workplace, the economics of poverty and social exclusion, and the householddivision of labor. In each case, the inclusion of identity substantively changesconclusions of previous economic analysis.

I. INTRODUCTION

This paper introduces identity—a person’s sense of self—intoeconomic analysis. Identity can account for many phenomena thatcurrent economics cannot well explain. It can comfortably resolve,for example, why some women oppose ‘‘women’s rights,’’ as seen inmicrocosm when Betty Friedan was ostracized by fellow suburban

* The authors especially wish to thank Abdeslam Maghraoui for his contin-ued help and insights and MichaelAsh, Jennifer Eichberger, and Cyd Fremmer forinvaluable research assistance. Henry Aaron, William Dickens, Claudia Goldin,Edward Glaeser, Lawrence Katz, Robert Merton, Anand Swamy, and an anony-mous referee made extensive comments on earlier drafts for which the authors areparticularly grateful. They also thank Robert Akerlof, Abhijit Banerjee, KaushikBasu, Paul Beaudry, Samuel Bowles, Robert Boyd, Gary Burtless, AlessandraCasella, Catherine Eckel, Stuart Elliott, Gary Fields, Pierre Fortin, James Foster,Richard Harris, Victoria Hattam, Peter Howett, Aurora Jackson, Kevin Lang,George Loewenstein, Glenn Loury, Michael Kremer, David Laibson, Janet Pack,Matthew Rabin, Francisco Rodriguez, Paul Romer, Eric Verhoogen, Eric Wanner,Kent Weaver, Robin Wells, Janet Yellen, and Peyton Young for help and comments.George Akerlof is grateful to the Canadian Institute for Advanced Research, theMacArthur Foundation, the Brookings Institution, and the National ScienceFoundation, under research grant number SBR 97-09250, for financial support.Rachel Kranton expresses her gratitude to the Russell Sage Foundation where shewas a Visiting Scholar for 1997–1998.

� 2000 by the President and Fellows of Harvard College and the Massachusetts Institute ofTechnology.The Quarterly Journal of Economics, August 2000

715

hou

sewives

forw

riting

Th

eF

emin

ine

Mystiqu

e.O

ther

problems

such

aseth

nic

and

racialcon

flict,

discrimin

ation,

intractable

labordispu

tes,and

separatistpolitics

allinvite

aniden

tity-basedan

alysis.Becau

seof

itsexplan

atorypow

er,num

erous

scholars

inpsych

ology,sociology,

politicalscien

ce,an

thropology,

and

history

have

adoptediden

tityas

acen

tralconcept.T

his

papersh

ows

how

identity

canbe

brough

tin

toecon

omic

analysis,

allowin

ga

new

viewofm

any

econom

icproblem

s. 1W

ein

corporateiden

tityin

toa

generalm

odelofbehavior

and

then

demon

stratehow

identity

influen

cesecon

omic

outcom

es.Specifi

cally,we

consider

gender

discrimin

ationin

the

laborm

ar-ket,

the

hou

sehold

divisionof

labor,an

dth

eecon

omics

ofsocial

exclusion

and

poverty.In

eachcase,

our

analysis

yieldspredic-

tions,

supported

byexistin

geviden

ce,th

atare

different

fromth

oseofexistin

gecon

omic

models.T

he

Con

clusion

indicates

man

yoth

errealm

sw

here

identity

almost

surely

matters.

Our

identity

model

ofbeh

aviorbegin

sw

ithsocial

difference.

Gen

der,auniversally

familiar

aspectofidentity,illu

strates.There

aretw

oabstract

socialcategories,

‘‘man

’’an

d‘‘w

oman

.’’T

hese

categoriesare

associatedw

ithdifferen

tideal

physical

attributes

and

prescribedbeh

aviors.Everyon

ein

the

population

isassign

eda

gender

category,as

either

a‘‘m

an’’

ora

‘‘wom

an.’’

Follow

ing

the

behavioral

prescriptions

foron

e’sgen

deraffirm

son

e’sself-

image,

oriden

tity,as

a‘‘m

an’’or

asa

‘‘wom

an.’’ 2

Violatin

gth

eprescription

sevokes

anxiety

and

discomfort

inon

eselfan

din

1.P

revious

econom

icliteratu

reon

identity

inclu

desF

olbre[1994]

who

discusses

the

importan

ceof

gender

identity

forcollective

actionth

atpreserves

male

privilege.O

ur

general

model

ofutility

allows

forth

isou

tcome,

asw

ellas

man

yoth

ersou

rcesof

gender

inequ

ality.Sen

[1985]m

ention

siden

tityas

anin

fluen

ceon

goalach

ievemen

t,bu

tdoes

not

incorporate

identity

into

autility

function

orm

odelsofspecifi

cecon

omic

settings.‘‘Iden

tity’’alsohas

other

connota-

tions:L

anda

[1994]and

Kevan

e[1994]con

siderhow

identity,defi

ned

asm

ember-

ship

ina

particular

group,affects

econom

ictran

sactions

when

individu

almem

bersare

subject

togrou

psan

ctions.

Bow

lesan

dG

intis

[1997]likew

isecon

sidercooperation

with

ina

comm

unity.

2.W

euse

the

word

prescriptionsrath

erth

ann

ormsbecau

sepreviou

susage

inecon

omics

has

giventh

elatter

termcon

notation

sth

atw

ould

bem

isleading

inth

econ

textof

this

paper.H

ere,agen

tsfollow

prescriptions,

forth

em

ostpart,

tom

aintain

their

self-concepts.

Incon

trast,in

much

ofth

eecon

omics

literature,

anorm

isobeyed

because

failure

todo

soresu

ltsin

punish

men

t(e.g.,A

kerlof[1976],K

andori

[1992],an

dC

ole,M

ailath,

and

Postlew

aite[1992]).

Oth

erau

thors,

how

ever,see

norm

sas

someth

ing

similar

toou

rprescription

s.In

Mon

tgomery’s

[1997]gam

e-theoretic

model

ofsocial

roles,agen

tsadopt

strategiesth

atnorm

sassign

their

rolesbecau

seoth

erwise

they

‘‘wou

ldnot

recognize

them

selves.’’Elster

[1989]writes

thatsocialn

orms

aresu

stained

bystron

gfeelin

gsofem

barrassmen

t,an

xiety,an

dgu

iltsuffered

fromviolatin

gth

em.

Huan

gan

dW

u[1994]

alsocon

sidersocial

norm

ssu

stained

bypeople’s

emotion

s,w

hich

inth

eview

ofth

ispaper

wou

ldresu

ltfrom

aperson

’ssen

seofself.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

716

others.G

ender

identity,th

en,ch

anges

the

‘‘payoffs’’fromdifferen

taction

s.T

his

modelin

gof

identity

isin

formed

bya

vastbody

ofresearch

onth

esalien

ceof

socialcategories

forhum

anbeh

avioran

din

teraction.

We

present

inth

enext

sectiona

seriesof

examples

ofidentity-related

behavior.T

hese

examples,an

doth

ereviden

ce,in

dicateth

at(1)

peoplehave

identity-based

payoffsderived

fromth

eirow

naction

s;(2)

peoplehave

identity-based

payoffsderived

fromoth

ers’actions;(3)th

irdparties

cangen

eratepersisten

tch

anges

inth

esepayoffs;

and

(4)som

epeople

may

choose

their

identity,bu

tch

oicem

aybe

proscribedfor

others.

The

concept

ofidentity

expands

econom

ican

alysisfor

atleast

four

correspondin

greason

s.F

irst,identity

canexplain

behavior

thatappears

detrimen

tal.P

eoplebeh

avein

ways

that

wou

ldbe

considered

maladaptive

oreven

self-destructive

byth

osew

ithoth

eriden

tities.The

reasonfor

this

behavior

may

beto

bolstera

sense

ofself

orto

salvea

dimin

ished

self-image.

Secon

d,iden

tityunderlies

anew

typeof

externality.

One

person’s

actions

canhave

mean

ing

foran

devoke

responses

inoth

ers.G

ender

againaffords

anexam

ple.Adress

isa

symbol

offem

inin

ity.Ifam

anw

earsa

dress,this

may

threaten

the

identity

ofoth

erm

en.

There

isan

externality,

and

furth

erextern

alitiesresu

ltifth

esem

enm

akesom

erespon

se.T

hird,

identity

revealsa

new

way

that

preferences

canbe

chan

ged.Notion

sof

identity

evolvew

ithin

asociety

and

some

inth

esociety

have

incen

tivesto

man

ipulate

them

.O

bvious

ex-am

plesoccu

rin

advertising

(e.g.,M

arlboroads).

As

we

shall

explore,th

ereare

man

yoth

ercases,

inclu

ding

public

policies,w

here

chan

ging

socialcategories

and

associatedprescription

saffects

econom

icou

tcomes.

Fou

rth,becau

seiden

tityis

fundam

entalto

behavior,ch

oiceof

identity

may

beth

em

ostim

portant

‘‘econom

ic’’decisionpeople

make.

Individu

alsm

ay—m

oreor

lesscon

sciously—

choose

who

they

wan

tto

be.L

imits

onth

isch

oicem

ayalso

beth

em

ostim

portant

determin

ant

ofan

individu

al’secon

omic

well-bein

g.P

revious

econom

ican

alysesof,for

example,poverty,labor

supply,

and

schoolin

ghave

not

considered

these

possibilities.O

ur

analysis

proceedsas

follows.

Inth

enext

sectionw

epropose

agen

eralutility

function

that

incorporates

identity

asa

motivation

forbeh

avior.Itintrodu

cesth

evocabu

laryan

dth

eoreti-cal

framew

orkused

throu

ghou

tth

epaper.

This

sectionalso

EC

ON

OM

ICS

AN

DID

EN

TIT

Y717

justifi

esou

rin

clusion

ofidentity

ina

utility

function

,presentin

ga

seriesof

examples

ofiden

tity-relatedbeh

avior.Section

IIIth

encon

structs

aprototype

game-th

eoreticm

odelof

identity

that

mirrors

standard

psychological

theory.

This

model

oftw

osocial

categories—G

reenan

dR

ed—con

tains

the

essential

elemen

tsof

socialdifferentiation

,identity,an

decon

omic

interaction

.Section

sIV,V,an

dV

Icon

sidergen

derdiscrim

ination

inth

elabor

market,

the

econom

icsof

povertyan

dsocial

exclusion

,and

the

hou

sehold

divisionoflabor,respectively.S

ectionV

IIcon

cludes

and

indicates

directions

forfu

ture

research.

II.U

TIL

ITY

FU

NC

TIO

NA

ND

EV

IDE

NC

EO

F

IDE

NT

ITY-R

EL

AT

ED

BE

HA

VIO

R

This

sectionproposes

autility

function

that

incorporates

identity

asa

motivation

forbeh

avior.We

drawon

extensive

work

inpsych

ologyan

ddiscu

ssspecifi

cexam

plesof

behavior

that

support

our

framew

ork.

A.A

Utility

Fu

nction

with

Iden

tity

Inou

rutility

function

,identity

isbased

onsocial

categories,C

.Each

personjh

asan

assignm

entofpeople

toth

esecategories,

cj ,so

that

eachperson

has

acon

ceptionofh

erow

ncategories

and

that

ofall

other

people. 3P

rescriptions

Pin

dicateth

ebeh

aviorappropriate

forpeople

indifferen

tsocial

categoriesin

different

situation

s.The

prescriptions

may

alsodescribe

anideal

foreach

categoryin

terms

ofphysicalch

aracteristicsan

doth

erattribu

tes.C

ategoriesm

ayalso

have

high

eror

lower

socialstatus.W

euse

the

word

identity

todescribe

botha

person’s

self-image

asw

ellasher

assigned

categories.G

ender

identity,

asin

dicatedearlier,

could

beform

alizedas

follows.T

here

isa

setofcategories

C,‘‘m

an’’an

d‘‘w

oman

,’’where

men

have

high

ersocial

status

than

wom

en.c

j describesj’s

own

gender

categoryas

wellas

j’sassign

men

tfor

everyone

elsein

the

population

.Passociates

toeach

categorybasic

physicalan

doth

erch

aracteristicsth

atcon

stitute

the

idealman

orw

oman

asw

ellasspecifi

esbeh

aviorin

differentsitu

ations

according

togen

der.E.g.,

the

idealwom

anis

female,th

in,an

dsh

ould

always

wear

adress;

3.A

nin

dividual

j’sm

apping

ofanoth

erin

dividualk

into

categoriesneed

not

correspond

tok’s

own

mappin

g.Inaddition

,socialcategoriesneed

not

bem

utu

allyexclu

sive,an

dan

individu

alm

aybe

mapped

into

severalsocial

categories(e.g.,

individu

aljis

botha

‘‘wom

an’’an

da

‘‘professional’’).

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

718

the

idealm

anis

male,m

uscu

lar,and

shou

ldnever

wear

adress,

exceptperh

apson

Hallow

een.

We

proposeth

efollow

ing

utility

function

:

(1)U

j �U

j (aj ,a

�j ,I

j ).

Utility

depends

onj’s

identity

orself-im

ageI

j ,asw

ellas

onth

eusu

alvectors

ofj’s

actions,

aj ,

and

others’action

s,a

�j .

Sin

cea

j

and

a�

j determin

ej’s

consu

mption

ofgoods

and

services,th

eseargu

men

tsan

dU

j (·)aresufficien

ttocaptu

reth

estan

dardecon

om-

icsofow

naction

san

dextern

alities.F

ollowin

gou

rdiscu

ssionabove,

we

proposeth

efollow

ing

representation

ofIj :

(2)I

j �I

j (aj ,a

�j ;c

j ,�j ,P

).

Aperson

j’siden

tityI

j depends,fi

rstof

all,onj’s

assigned

socialcategories

cj .

The

socialstatu

sof

acategory

isgiven

byth

efu

nction

Ij (·),an

da

personassign

eda

categoryw

ithhigh

ersocial

status

may

enjoy

anen

han

cedself-im

age.Iden

tityfu

rther

de-pen

dson

the

extent

tow

hich

j’sow

ngiven

characteristics

�j

match

the

idealofj’s

assigned

category,indicated

byth

eprescrip-

tions

P. 4

Fin

ally,identity

depends

onth

eexten

tto

which

j’sow

nan

doth

ers’actions

correspond

toprescribed

behavior

indicated

byP

.W

ecall

increases

ordecreases

inutility

that

derivefrom

Ij ,

gains

orlosses

inid

entity. 5

Inth

esim

plestcase,

anin

dividual

jch

oosesaction

sto

maxim

izeutility

(1),taking

asgiven

cj ,�

j ,and

Pan

dth

eaction

sof

others.W

euse

the

verb‘‘ch

oose’’advisedly.We

donotpresu

me

one

way

oran

other

that

peopleare

aware

oftheir

own

motivation

s,asin

standard

utility

theory

which

isagn

osticas

tow

heth

eran

individu

alshopper

isaw

areor

not

ofthe

reasons

forher

choices. 6

Beyon

daction

s,tosom

eexten

tan

individu

almay

alsoch

ooseth

ecategory

assignm

ent

cj .S

ocialcategoriesm

aybe

more

orless

ascriptive,an

din

general,

the

individu

alis

likelyto

have

some

4.In

the

caseofa

categoryw

ithhigh

(low)socialstatu

s,aperson

jmay

gainw

hen

own

characteristics

areclose

to(far

from)from

the

ideal.5.

Sin

cean

individu

al’sself-con

ceptm

aybe

formed

byseein

gon

eselfthrou

ghth

eeyes

ofothers

[Gleitm

an1996,p.343],th

esegain

sor

lossesm

ayalso

depend

onhow

others

interpret

i’saction

s.The

opinion

sof

others

may

berevealed

throu

ghaction

sa

�j ;th

ein

dividualm

ayalso

careabou

toth

ers’categorizations

c�

j .6.

Sen

[1997]m

akesth

ean

alogyth

atligh

tdoes

not

know

that

itis

min

imizin

gdistan

ce,bu

tbeh

avesas

ifit

does.T

his

notion

follows

Friedm

an’s

[1953]dicta

forth

em

ethodology

ofpositive

econom

ics.W

heth

eror

not

jcon

-sciou

slyrealizes

she

ism

aximizin

ga

utility

function

such

as(1),

she

doesso

neverth

eless.In

our

setting,

inparticu

lar,th

em

otivations

forbeh

aviorm

aybe

uncon

scious.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y719

choice

overiden

tity,asin

deedpeople

may

evenhave

some

choice

overth

eirgen

der.A

gain,

this

‘‘choice’’

may

bem

oreor

lesscon

scious.

Individu

alactions

may

alsoaffect

the

prescriptions

P,th

eset

ofsocialcategories

C,as

wellas

the

status

ofdifferen

tcategories

reflected

inI

j (·).W

ithrespect

togen

der,for

example,

status

differences

between

men

and

wom

enhave

dimin

ished

overtim

e,an

dprescribed

behavior

and

physical

idealshave

chan

ged.Gen

-der

categoriesth

emselves

have

become

variedan

dcom

plex.There

may

beno

universal

agreemen

tabou

tsocial

categoriesan

dprescription

s.In

deed,th

eyare

the

subject

ofm

uch

debatean

dcon

troversy.

B.P

sychology

and

Experim

ents

onG

roup

Iden

tification

The

promin

ence

ofiden

tityin

psychology

suggests

that

econom

istssh

ould

consider

identity

asan

argum

ent

inutility

function

s.P

sychologists

have

long

positeda

selfor

‘‘ego’’as

aprim

aryforce

ofindividu

albehavior.T

hey

have

furth

erassociated

anin

dividual’s

sense

ofselftoth

esocialsettin

g;identity

isbou

nd

tosocial

categories;and

individu

alsiden

tifyw

ithpeople

insom

ecategories

and

differentiate

them

selvesfrom

those

inoth

ers. 7W

hile

experimen

tsin

socialpsych

ologydo

not

show

the

existence

ofa

‘‘self’’orth

isiden

tification

perse,

they

dodem

on-

strateth

ateven

arbitrarysocial

categorizations

affectbeh

avior. 8C

onsider

the

Robbers

Cave

experimen

t.In

itsin

itialw

eek,tw

ogrou

psof

boysat

asu

mm

ercam

pin

Oklah

oma

were

keptapart.

Durin

gth

isperiod,

the

boysdeveloped

norm

sof

behavior

and

identities

asbelon

ging

toth

eirgrou

p.W

hen

they

met

fora

tourn

amen

tin

the

second

week,th

eeleven

-year-oldequ

ivalentof

war

brokeou

t,w

ithnam

e-calling,

stereotyping,

and

figh

ting.

Later

experimen

tssh

owth

atcom

petitionis

not

necessary

forgrou

piden

tification

and

eventh

em

ostmin

imalgrou

passign

men

tcan

affectbeh

avior.‘‘Grou

ps’’formby

noth

ing

more

than

random

assignm

entofsu

bjectsto

labels,such

aseven

orodd.S

ubjects

arem

orelikely

togive

rewards

toth

osew

ithth

esam

elabel

than

toth

osew

ithoth

erlabels,

evenw

hen

choices

arean

onym

ous

and

have

no

impacton

own

payoffs.Subjects

alsohave

high

eropin

ions

ofmem

bersofth

eirow

ngrou

p.

7.F

ordiscu

ssionofth

e‘‘self,’’see

Thom

as[1996],B

reger[1974],or

Gleitm

an[1996].

For

areview

ofth

esocial

psychology

ofiden

tity,see

Brow

n[1986]

and

Weth

erell[1996],and

especiallyth

ew

orkofTajfelan

dTu

rner

[1979].8.

For

discussion

ofsocial

psychology

experimen

ts,see

Brow

n[1986,

pp.541–566]an

dW

etherell[1996,pp.203–216].

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

720

Our

modelin

gofiden

tityexactly

parallelsth

eseexperim

ents.

Inth

eexperim

ents,as

inou

rutility

function

(1),there

aresocial

categories;th

ereis

anassign

men

tof

subjects

toth

osesocial

categories;finally,su

bjectshave

inm

ind

some

formofassign

men

t-related

prescriptions,

elserew

ardsw

ould

not

depend

ongrou

passign

men

t.

C.E

xamples

ofIden

tity-Related

Beh

avior

We

next

present

aset

of‘‘real-w

orld’’exam

plesof

four

differentw

ays,outlin

edin

the

introdu

ctionan

dform

alizedin

our

utility

function

,that

identity

may

influen

cebeh

avior.O

ur

first

setdem

onstrates

that

peoplehave

identity-related

payoffsfrom

their

own

actions.

The

impact

ofan

actiona

jon

utility

Uj depen

dsin

parton

itseffect

oniden

tityI

j .

Self-M

utilation

.T

he

first

ofth

eseexam

plesis

perhaps

the

most

dramatic:

peoplem

utilate

their

own

orth

eirch

ildren’s

bodiesas

anexpression

ofiden

tity.Tattooing,body-piercin

g(ear,

nose,

navel,

etc.),hair

conkin

g,self-starvation

,steroid

abuse,

plasticsu

rgery,an

dm

alean

dfem

alecircu

mcision

allyield

physical

markers

ofbelon

ging

tom

oreor

lessexplicit

socialcategories

and

groups. 9

Interm

sof

our

utility

function

,th

esepractices

transform

anin

dividual’s

physical

characteristics

tom

atchan

ideal. 10T

he

mutilation

may

occur

because

peoplebelieve

itleads

topecu

niary

rewards

and

interaction

ssu

chas

marriage.B

utth

eten

acityan

ddefen

seofth

esepractices

indicate

the

extent

tow

hich

belongin

grelies

onritu

al,an

dpeople

have

intern

alizedm

easures

ofbeauty

and

virtue. 11

Gen

der

and

Occu

pations.

Fem

aletrial

lawyer,

male

nurse,

wom

anM

arine—

allcon

jure

contradiction

s.W

hy?

Becau

setrial

9.See

Khatibi

[1986]for

analysis

ofhow

markin

gth

ebody,by

circum

cisionan

dtribaltattoos,m

arksth

eself.

10.A

naltern

ativeexplan

ationis

that

these

practicesare

signals

ofsom

eunobserved

econom

icallyrelevan

tattribu

te.H

owever,

itis

hard

toim

agine

why

individu

alcostsofth

esesign

alsw

ould

becorrelated

with

these

attributes.

11.In

astu

dyofsexu

alityin

ruralE

gypt,Khattab

[1996]reportsth

atw

omen

consider

female

circum

cisiona

beautifyin

gpractice.It

accentu

atesth

edifferen

cebetw

eenth

esexes:‘‘W

edon

’tw

ant

tolook

likea

man

with

aprotru

ding

organ’’[p.

20].Bum

iller[1990]

reportsan

example

offem

aledefen

seof

female

self-sacrifice.

Both

men

and

wom

enjou

rneyed

topay

their

respectsafter

ayou

ng

wom

ancom

mitted

satiin

aR

ajasthan

ivillage

in1987.

Sati

isth

epractice

ofth

ew

idowbu

rnin

gto

deathon

her

husban

d’sfu

neral

pyre.O

ne

devoteeexpressed

her

admiration

:‘‘IfIhad

know

nsh

ew

asgoin

gto

doth

isIw

ould

have

touch

edher

feet.N

owIw

illgiveher

aplace

inm

yhou

sean

dw

orship

her

everyday.’’T

his

respectis

no

lessdim

inish

edby

admirers’dou

btsth

atthey

wou

ldhave

had

the

same

courage

orby

their

ignoran

ceofth

epressu

reon

the

widow

fromher

in-law

s.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y721

lawyers

areview

edas

mascu

line,

nurses

asfem

inin

e,an

da

Marin

eas

the

ultim

atem

an.P

eoplein

these

occupation

sbu

tof

the

oppositesex

oftenhave

ambigu

ous

feelings

about

their

work.

Interm

sof

our

utility

function

,an

individu

al’saction

sdo

not

correspond

togen

derprescription

sof

behavior.A

revealing

study

inth

isregard

isP

ierce’s[1995]

participant-observer

researchon

the

legalprofession

. 12F

emale

lawyers

thou

ght

ofth

emselves

asw

omen

,yetbeing

agood

lawyer

mean

tacting

likea

man

.Law

yersw

eretold

intrain

ing

sessions

toact

like‘‘R

ambo’’an

dto

‘‘takeno

prisoners.’’In

the

office,trialattorneys

who

didnot

‘‘win

big’’were

describedas

‘‘havin

gno

balls.’’In

timidation

ofw

itnesses

was

‘‘mach

oblasts

against

the

other

side.’’AC

hristm

asskit

abouttw

opartn

ersdram

atizedth

egen

dercon

flict:

[O]n

esecretary

dressedup

asR

achel

and

anoth

erdressed

up

asM

ichael.T

he

secretaryportrayin

gM

ichael...ran

around

the

stagebarkin

gorders

and

singin

g,‘‘I’mM

ichaelB

ond,I’m

such

abu

sym

an.I’m

such

abu

sym

an.’’T

he

other

secretaryfollow

edsu

itby

barking

ordersan

dsin

ging,‘‘I’m

Rach

elR

osen,I’m

such

abu

sym

an,I

mean

wom

an.I’m

such

abu

sym

an,I

mean

wom

an....’’M

ichael

responded

toth

espoof

instride....

Rach

el,on

the

other

han

d,was

veryupset

[Pierce,1995,p.130].

Fem

alelaw

yersexpressed

their

ambivalen

cein

man

ydiscu

s-sion

s.‘‘C

andace,’’an

other

partner,

toldP

ierce:‘‘I

had

forgottenhow

much

anger

I’vebu

riedover

the

yearsabou

tw

hat

happen

edto

the

wom

anw

ho

became

alaw

yer....To

bea

lawyer,

some-

where

along

the

way,I

made

adecision

that

itm

eantactin

glike

am

an.To

doth

atIsqu

eezedth

efem

alepart

ofme

into

abox,pu

ton

the

lid,and

tucked

itaw

ay’’[Pierce

1995,p.134].

Alu

mn

iG

iving.

Charitable

contribu

tions

may

yielda

‘‘warm

glow’’[A

ndreon

i1989],buthow

dopeople

choose

one

organization

overan

other?

Charity

toth

eorgan

izationw

ithth

ehigh

estm

arginalretu

rnw

ould

maxim

izeits

econom

icim

pact.Yet,atleastfor

high

eredu

cation,

contribu

tions

may

well

reflect

identity.

Gradu

atesgive

toth

eirow

nalm

am

ater.A

lum

ni

giving

could

enhan

ceth

evalu

eof

adegree

bym

aintain

ing

anin

stitution

’srepu

tation.B

utth

isexplan

ationsuffers

fromth

ecollective

actionproblem

.And

itdoesnotaccou

ntfor

studen

tloyaltyan

diden

tifica-

tionw

ithan

institu

tion,as

expressedin

such

lyricsas

‘‘For

God,

forcou

ntry,an

dfor

Yale.’’

12.F

ora

study

ofnurses

and

Marin

es,seeW

illiams

[1989].

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

722

Mou

ntain

eering.

Why

dopeople

climb

mou

ntain

s?L

oewen

-stein

[1998]argues

thatfacin

gth

eextrem

ediscom

fortand

danger

ofmou

ntain

eering

enhan

cesan

individu

al’ssen

seofself.

Our

second

setof

examples

demon

stratesth

atpeople

have

identity-related

payoffsfrom

others’

actions.

The

effectof

anaction

a�

j onutility

inclu

desan

impact

onI

j .

Gen

der

and

Occu

pations.

Aw

oman

workin

gin

a‘‘m

an’s’’job

may

make

male

colleagues

feelless

like‘‘m

en.’’To

allayth

esefeelin

gs,they

may

actto

affirmth

eirm

asculin

ityan

dact

against

female

coworkers.In

her

study

ofcoalh

andlers

ina

power

plant,

Padavic

[1991]interpreted

the

behavior

ofher

male

coworkers

inth

isw

ay.On

one

occasion,th

eypicked

her

up,tossed

her

backan

dforth

,an

dattem

ptedto

push

her

onto

the

coalcon

veyerbelt

(jokingly,ofcou

rse).Inth

ecase

ofanoth

erw

orker,no

one

trained

her,

no

one

helped

her,

and

when

she

askedfor

help,

she

was

refused

assistance

that

wou

ldhave

beenrou

tine

form

alecow

orkers. 13To

furth

erassay

the

reasons

forsu

chbeh

avior,w

etook

aran

dom-sam

pleteleph

one

survey

relating

avign

etteabou

ta

female

carpenter

ata

constru

ctioncom

pany

who

was

‘‘baitedan

dteased’’by

am

alecow

orker.We

seein

TableI

that

amon

gth

esix

possibleexplan

ations,

84percen

tof

the

responden

tssaid

itw

as‘‘som

ewhat

likely,’’‘‘likely,’’or‘‘very

likely’’that

the

male

worker

behaved

inth

isw

aybecau

sehe

feltless

mascu

line. 14

This

explanation

was

one

ofth

em

ostpopu

lar,an

dm

oreth

anth

ree-qu

artersof

the

responden

tsth

ough

tth

ata

wom

anin

am

an’s

job‘‘frequ

ently’’or

‘‘almost

always’’faces

such

treatmen

t.

Man

hood

and

Insu

lt.F

ora

man

,anaction

may

beview

edas

anin

sult

which

,ifleft

unan

swered,im

pugn

shis

mascu

linity.A

sin

the

example

above,an

actiona

�j

impacts

Ij

which

may

becou

ntered

byan

actiona

j .Psych

ologistsN

isbettan

dC

ohn

[1996]have

detectedsu

chiden

titycon

cerns

inexperim

ents

atth

e

13.L

evine

[1997]also

found

that

men

oftenrefu

sedto

trainw

omen

and

sabotagedth

eirw

ork.In

addition,

wom

enin

men

’sjobs

were

subject

tosexu

alin

nuen

do.For

acollection

ofsuch

examples

seeSch

ultz

[1998].14.

Differen

cesin

response

bygen

derw

erenegligible.

The

survey

inclu

dedth

reeoth

ervign

ettes,tw

oof

which

describeda

man

(wom

an)

contem

plating

asw

itchto

apredom

inan

tlyfem

ale(m

ale)occu

pation.

Respon

sesin

dicateth

atgen

dercou

ldbe

ofcon

cernin

such

adecision

.The

responses

were

unin

formative,

how

ever,when

the

switch

was

otherw

iseundesirable

soth

atan

ygen

derconfl

ictw

ould

bem

oot.R

esponses

toth

elast

vignette

strongly

suggest

that

identity

consideration

sare

am

ajorreason

fortakin

gth

etim

eto

vote.Our

sample

was

half

male,h

alffemale,an

d60

percentcollege

graduates.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y723

University

ofM

ichigan

.T

hese

experimen

ts,th

eyargu

e,reveal

remnan

tsof

the

white

antebellu

mSou

thern

‘‘cultu

reof

hon

or’’indisparate

reactions

toin

sult

ofm

alesfrom

the

U.S

.Sou

than

dN

orth. 15

Their

experimen

tsin

volvedvariation

sof

the

followin

g

15.F

ora

descriptionof

this

‘‘cultu

reof

hon

or,’’seealso

Butterfi

eld[1995].

‘‘Gen

tlemen

’’reactedto

insu

ltby

engagin

gin

duels.

Those

oflow

erclass

fough

tw

ithhan

dsan

dfists

with

no

holds

barred,so

that

figh

tsexten

dedto

such

extremities

aseyes,ears,an

dnose.

TA

BL

EI

VIG

NE

TT

EC

ON

CE

RN

ING

HA

RA

SSM

EN

TA

ND

EV

AL

UA

TIO

NO

FP

OSSIB

LE

EX

PL

AN

AT

ION

S

Vign

ette:Pau

lisa

carpenter

fora

constru

ctioncom

pany.T

he

compan

yhas

just

hired

Christin

e,itsfirst

female

carpenter,for

3dollars

lessper

hou

rth

anit

paysP

aulan

dth

eoth

ercarpen

ters.On

Christin

e’sfirst

dayofw

ork,Pau

land

two

ofhis

coworkers

baitan

dtease

Christin

e,makin

git

difficult

forher

todo

her

job.Try

toim

agine

why

Pau

lbehaved

ashe

did.Rate

eachofth

efollow

ing

explana-

tions

forP

aul’s

behavior

asnot-at-alllikely,n

otlikely,som

ewhat

likely,likely,or

verylikely.

Explan

ation

Fraction

somew

hat

likely,likely,or

verylikely

a,bAveragescore

c

Pau

lputC

hristin

edow

nbecau

sehe

isafraid

that

byhirin

ga

wom

anth

ecom

pany

canlow

erhis

wage.

.36(.06)

2.5(.12)

Pau

lputC

hristin

edow

nbecau

sehe

doesnot

feelthat

itis

fairth

atC

hristin

eis

getting

alow

erw

age..13

(.04)1.7

(.12)P

aulpu

tC

hristin

edow

nbecau

sehe

feelsless

mascu

-lin

ew

hen

aw

oman

isdoin

gth

esam

ejob.

.84(.04)

3.4(.12)

Pau

lputC

hristin

edow

nbecau

sehe

feelshe

and

his

friends

willn

otbe

ableto

jokearou

nd

ifaw

oman

ispresen

t.

.84(.04)

3.6(.12)

Pau

lputC

hristin

edow

nbecau

sehe

isafraid

that

other

men

willtease

him

ifaw

oman

isdoin

gth

esam

ejob.

.76(.05)

3.3(.13)

Pau

lputC

hristin

edow

nbecau

sehe

isafraid

that

peoplew

illthin

kth

athis

jobrequ

iresless

skillifaw

oman

isdoin

gth

esam

ejob.

.64(.06)

2.9(.12)

Pau

lputC

hristin

edow

nbecau

sehe

isafraid

that

ifhe

doesnot,th

enhis

male

coworkers

willstart

totease

him

.

.80(.05)

3.4(.13)

Pau

lputC

hristin

edow

nbecau

sehe

feelsth

atit

isw

rong

forw

omen

tow

orkin

am

an’s

job..77

(.05)3.2

(.14)

a.Sam

plesize

is70

hou

seholds.H

ouseh

oldsw

ereselected

random

lyfrom

the

Frem

ont,C

Aph

onebook.

b.Stan

darderrors

arein

parenth

eses.c.A

veragew

ithnot-at-alllikely

�1,n

otlikely

�2,som

ewhat

likely�

3,likely�

4,verylikely

�5.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

724

scenario:

anassociate

ofth

eexperim

enters

bum

pedsu

bjectsin

the

hallw

ayas

they

made

their

way

toth

eexperim

ent.

Rath

erth

anapologizin

g,th

eassociate

calledth

esu

bject‘‘assh

ole.’’In-

sulted

Sou

thern

ersw

erem

orelikely

than

insu

ltedN

orthern

ersan

dcon

trolSou

thern

ersto

fill

insu

bsequen

tw

ord-completion

testsw

ithaggressive

words

(forexam

ple,g-un

rather

than

f-un)

and

had

raisedcortisollevels.

Most

revealing

that

the

insu

ltaffected

identity,

insu

ltedSou

thern

ersw

erealso

more

likelyto

fearth

atth

eexperim

enter

had

alow

opinion

ofth

eirm

asculin

ity.T

hey

will

probablynever

meet

the

experimen

teror

the

hallw

ayaccom

pliceagain

;th

eiren

counter

inth

eexperim

ent

isoth

erwise

anon

ymou

s.T

heir

concern

about

the

experimen

terth

encan

only

bea

concern

about

how

they

feelaboutth

emselves,abou

tth

eirow

nsen

seofiden

tity,as

perceivedth

rough

the

‘‘mirror

ofthe

opinion

san

dexpectation

sofoth

ers’’[Gleitm

an1996,p.343].W

esee

the

same

psychology

inoth

erexam

ples.

Ch

angin

gG

roups

orV

iolating

Prescription

s.B

ecause

ofj’s

iden

tification

with

others,it

may

affectj’s

identity

when

anoth

erperson

inj’s

socialcategory

violatesprescription

sor

becomes

adifferen

tperson

. 16A

comm

onrespon

seis

scornan

dostracism

,w

hich

distances

oneself

fromth

em

averickan

daffirm

son

e’sow

nself-im

age.Such

behavior

occurs

dailyin

school

playgrounds,

where

children

who

behave

differently

arem

ockedan

dtau

nted.

Those

who

seekupw

ardm

obilityare

oftenteased

byth

eirpeers,

asin

AH

opein

the

Un

seen[S

uskin

d1998],w

hich

describesC

edricJen

nin

gs’progressfrom

one

ofW

ashin

gton’s

most

blighted

high

schools

toB

rown

University.

The

bookopen

sw

ithC

edricin

the

high

-schoolch

emistry

lab,escaping

the

catcallsofth

ecrow

dat

anaw

ardsassem

bly.T

hose

who

tryto

chan

gesocial

categoriesan

dprescription

sm

ayface

similar

derisionbecau

seth

ech

ange

may

devalue

others’iden

tity,asfor

the

hou

sewives

inB

ettyF

riedan’s

subu

rb.

Our

third

setof

examples

demon

stratesth

atto

some

extent

peoplech

ooseth

eiriden

tity;that

is,cj m

aybe

partiallya

choice.

Man

yw

omen

inth

eU

nited

States

canch

ooseeith

erto

bea

careerw

oman

ora

hou

sewife

(seeG

erson[1986]).P

arents

oftench

oosea

school—

public

versus

private,secular

versus

parochial—

toin

flu-

16.W

ediscu

ssth

epsych

ologyofid

entifi

cationan

dits

implication

sfu

rther

inth

enext

section.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y725

ence

ach

ild’sself-im

age,iden

tification

with

others,

and

behav-

ior. 17T

he

choice

ofw

here

tolive

atcollege

canboth

reflect

and

chan

gehow

studen

tsth

ink

ofthem

selves.Fratern

ities,sororities,A

frican-A

merican

,oroth

er‘‘th

eme’’-orien

teddorm

sare

allassoci-ated

with

socialgrou

ps,self-im

ages,an

dprescribed

behavior. 18

The

listcan

contin

ue.

The

choice

foran

imm

igrant

tobecom

ea

citizenis

noton

lya

chan

gein

legalstatus

buta

chan

gein

identity.

The

decisionis

thus

oftenfrau

ghtw

itham

bivalence,an

xiety,and

evengu

ilt.Iden

tity‘‘ch

oice,’’how

ever,is

veryoften

limited.

Ina

societyw

ithracialan

deth

nic

categories,forexam

ple,those

with

non

dis-tin

guish

ing

physicalfeatu

resm

aybe

ableto

‘‘pass’’asa

mem

berof

anoth

ergrou

p.B

ut

others

will

becon

strained

byth

eirappear-

ance,voice,or

accent.

Our

fourth

setof

examples

demon

stratesth

ecreation

and

man

ipulation

ofsocialcategoriesC

and

prescriptions

P. 19

Ad

vertising.

Advertisin

gis

anobviou

sattem

ptto

man

ipu-

lateprescription

s.M

arlboroan

dV

irginia

Slim

sadvertisem

ents,

forexam

ple,prom

otean

image

ofth

eideal

man

orw

oman

complete

with

the

rightcigarette. 20

Profession

alan

dG

radu

ateS

chools.

Gradu

atean

dprofes-

sionalprogram

stry

tom

oldstu

dents’beh

aviorth

rough

ach

ange

iniden

tity.As

a‘‘on

e-L’’H

arvardL

awSch

oolstuden

tsaid:‘‘‘T

hey

aretu

rnin

gm

ein

tosom

eone

else.They’re

makin

gm

edifferen

t’’’[Tu

row1977,

p.73].

Inm

edicine,

theology,

the

military,

and

the

doctorate,a

titleis

addedto

agradu

ate’snam

e,su

ggesting

the

chan

gein

person.

Political

Iden

tity.P

oliticsis

oftena

battleover

identity. 21

Rath

erth

antake

preferences

asgiven

,political

leadersan

d

17.C

atholic

schools

inth

eU

nited

States

atth

een

dofth

enin

eteenth

centu

ryw

erea

bridgebetw

eenim

migran

ts’old

European

identities

and

their

new

Am

ericanselves

[Bryk,

Lee,

and

Hollan

d1993,

p.27].

Muslim

schools,

whose

enrollm

entis

curren

tlygrow

ing,are

partlyrefu

gesfrom

public

schoolsystem

s,but

parents

alsoch

ooseth

emto

instillin

their

children

aM

uslim

identity

and

respectfor

behavioralprescription

s,and

tocou

nter

what

man

yview

asa

distortedim

ageofM

uslim

san

dIslam

inA

merica

[Sach

s1998].

18.F

oran

anth

ropologicalstudy

ofidentity,fratern

ities,and

prescriptions

forbroth

ers’behavior,see

San

day[1990].

19.T

he

socialevolu

tionan

dcon

struction

ofgrou

pdistin

ctions

and

socialcategories

isth

esu

bjectof

much

research.F

ora

survey,see

Weth

erell[1996,pp.

219–227].20.

See

deG

razia’s[1996]

volum

efor

historical

studies

ofadvertisin

gan

doth

erin

fluen

ceson

gender

and

consu

mption

.21.

For

theory

and

analysis

ofpoliticalidentity,see

Norton

[1988].

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

726

activistsoften

striveto

chan

gea

population

’spreferen

cesth

rough

ach

ange

iniden

tityor

prescriptions. 22

Again

,exam

plesabou

nd.

Fascist

and

populist

leadersare

infam

ous

forth

eirrh

etoricfosterin

gracial

and

ethnic

divisions,

with

tragiccon

sequen

ces.Sym

bolicacts

and

transform

ediden

titiesspu

rrevolu

tions.

The

ringin

gof

the

Liberty

Bell

calledon

the

colonists’iden

titiesas

Am

ericans.

Gan

dhi’s

Salt

March

sparkedan

Indian

nation

aliden

tity.T

he

Fren

chR

evolution

chan

gedsu

bjectsin

tocitizen

s,an

dth

eR

ussian

Revolu

tiontu

rned

them

into

comrad

es.

III.E

CO

NO

MIC

SA

ND

IDE

NT

ITY:A

PR

OT

OT

YP

EM

OD

EL

Inth

issection

we

constru

cta

prototypem

odelof

econom

icin

teractionin

aw

orldw

here

identity

isbased

onsocialdifferen

ce.In

additionto

the

usu

altastes,

utility

fromaction

sw

illalso

depend

oniden

tity.Identity

willdepen

don

two

socialcategories—G

reenan

dR

ed—an

dth

ecorrespon

dence

ofow

nan

doth

ers’action

sto

behavioralprescription

sfor

their

category.

A.A

Prototype

Mod

el

We

beginw

ithstan

dardecon

omic

motivation

sfor

behavior.

There

aretw

opossible

activities,Activity

One

and

Activity

Two.

There

isa

population

ofin

dividuals

eachof

whom

has

ataste

foreith

erA

ctivityO

ne

orTw

o.If

aperson

with

ataste

forA

ctivityO

ne

(Two)

undertakes

Activity

One

(Two),sh

eearn

sutility

V.A

nin

dividualw

ho

chooses

the

activityth

atdoes

not

match

her

tasteearn

szero

utility.

Ina

standard

model

ofutility

maxim

ization,

eachperson

wou

lden

gagein

the

activitycorrespon

ding

toher

taste.We

next

constru

ctiden

tity-basedpreferen

ces.W

esu

pposeth

atth

ereare

two

socialcategories,

Green

and

Red.

We

assum

eth

esim

plestdivision

ofthe

population

into

categories;allpersons

thin

kof

them

selvesan

doth

ersas

Green

. 23W

eadd

simple

behavioral

prescriptions:

aG

reensh

ould

engage

inA

ctivityO

ne

(incon

trastto

Reds

who

engage

inA

ctivityTw

o).A

nyon

ew

ho

chooses

Activity

Two

isnot

a‘‘tru

e’’Green

—sh

ew

ould

loseher

22.R

omer

[1994]has

considered

the

possibilityth

atpolitician

scan

man

ipu-

latevoters’

emotion

s,in

particular

their

‘‘anger,’’

and

thereby

affectpolitical

outcom

es.23.

Ofcou

rse,itis

possibleth

atnot

everyone

thin

ksof

herself

asG

reen.W

ediscu

ssth

epossibility

ofdifferen

tiden

titiesan

doth

erexten

sions

toth

em

odelbelow

.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y727

Green

identity.T

his

lossin

identity

entails

aredu

ctionin

utility

ofI

s ,w

here

the

subscript

sstan

dsfor

‘‘self.’’Inaddition

,th

ereare

identity

externalities.Ifan

iand

jarepaired,A

ctivityTw

oon

the

partof

idim

inish

esj’s

Green

identity.

jhas

aloss

inutility

Io ,

where

the

subscript

oden

otes‘‘oth

er.’’A

fteri

has

comm

ittedA

ctivityTw

o,jm

ay‘‘respon

d.’’The

response

restoresj’s

identity

ata

costc,w

hile

entailin

ga

lossto

iinam

ount

L. 24

Figu

reI

represents

anin

teractionbetw

eenan

individu

alw

itha

tastefor

Activity

One

(‘‘Person

One’’)

and

anin

dividual

with

ataste

forA

ctivityTw

o(‘‘P

ersonTw

o’’).Person

One

chooses

anactivity

first. 25

This

model

canbe

expressedby

ideascen

tralto

the

psycho-

dynam

icth

eoryof

personality,

found

inalm

ostan

ypsych

ologytext. 26

Inperson

alitydevelopm

ent,

psychologists

agreeon

the

importan

ceof

intern

alizationof

rules

forbeh

avior.F

reud

calledth

isprocess

the

developmen

tof

the

superego.

Modern

scholars

disagreew

ithF

reud

onth

eim

portance

ofpsychosexu

alfactorsin

anin

dividual’s

developmen

t,but

they

agreeon

the

importan

ceof

anxiety

that

aperson

experiences

when

she

violatesher

intern

al-ized

rules.

One’s

iden

tity,or

ego,or

self,m

ust

becon

stantly

‘‘defended

against

anxiety

inorder

tolim

itdisru

ptionan

dm

ain-

taina

sense

ofunity’’[T

hom

as1996,p.284].In

terms

ofourm

odel,P

ersonTw

o’sin

ternalization

ofprescriptions

causes

her

tosuffer

aloss

inutility

ofI

sif

she

chooses

Activity

Two.

Toavoid

this

anxiety,sh

em

ayrefrain

fromth

atactivity.

Iden

tification

isa

criticalpartofth

isin

ternalization

process:a

personlearn

sa

setof

values

(prescriptions)

such

that

her

actions

shou

ldcon

formw

ithth

ebeh

aviorof

some

peoplean

dcon

trastw

ithth

atof

others.

IfP

ersonO

ne

has

intern

alizedprescription

svia

such

identifi

cations,

anoth

erperson

’sviolation

ofth

eprescription

sw

illcau

sean

xietyfor

Person

One. 27

Inou

rm

odel,th

isan

xietyis

modeled

asa

lossin

utility

ofI

o .P

ersonO

ne’s

response,

inou

rlan

guage,

restoresher

identity,

and

interm

sof

the

psychology

textbookrelieves

her

anxiety

and

main

-

24.In

Rabin

’s[1993]th

eoryoffairn

ess,agents

arew

illing

topay

tobe

‘‘mean

’’to

those

who

are‘‘m

ean’’to

them

.The

similarity

isprobably

no

coinciden

ce.Alikely

reasonfor

such

arespon

seis

preservationofself-im

age.25.

Sin

ceP

ersonO

ne

never

chooses

Activity

Two

ina

subgam

eperfect

equilibriu

m,w

esu

ppressth

isbran

chofth

etree.

26.See,

forexam

ple,G

leitman

[1996,C

hapter

17],T

hom

as[1996],

and

Breger

[1974].27.

The

violationarou

sesem

otions

that

Person

One

has

repressedin

the

processof

intern

alizing

the

behavioral

rules.

The

psychoan

alyticth

eory,th

en,

suggests

uncon

scious

motivation

sfor

behavior.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

728

FIG

UR

EI

Gam

eTree

ofInteraction

between

Person

One

and

Person

Two

EC

ON

OM

ICS

AN

DID

EN

TIT

Y729

tains

her

sense

ofunity.P

ersonO

ne

no

longer

losesI

o ,althou

ghsh

edoes

incu

rc. 28

B.E

quilibriu

mO

utcom

es

There

arefou

rpossible

subgam

eperfectou

tcomes

ofthe

game

inF

igure

I.(i)

Person

One

detersP

ersonTw

ofrom

engagin

gin

Activity

Two,w

hen

c�

Io an

dI

s�

V�

Is

�L

.(ii)

Person

One

responds

but

doesnot

deterP

ersonTw

ofrom

engagin

gin

Activity

Two,w

hen

c�

Io an

dI

s�

L�

V.

(iii)Person

One

doesnot

respond,an

dP

ersonTw

oen

gagesin

Activity

Two,w

hen

c�

Io an

dI

s�

V.

(iv)Person

Two

doesnot

engage

inA

ctivityTw

oregardless

ofP

ersonO

ne’s

response,w

hen

Is

�V

.T

his

simple

model

affordsth

reelesson

s.F

irst,as

discussed

earlier,th

em

odelestablish

esth

econ

nection

between

econom

icin

teractions

and

the

psychology

ofidentity,especially

the

implica-

tions

ofid

entifi

cation.

Secon

d,th

em

odelallow

sa

comparative

staticanalysis

onidentity-related

parameters.F

inally,theelem

en-tary

assum

ptions

ofth

em

odelsu

ggestexten

sions

that

entail

greaterrealism

and

furth

erim

plications

ofiden

tityfor

econom

icin

teraction.

C.C

omparative

Statics

Com

parativestatics

show

how

traditional

econom

icpolicies

canaffect

behavior

inth

issettin

g.For

example,a

‘‘tax’’Ton

the

response

toA

ctivityTw

ow

illaffect

the

equilibriu

mou

tcome

incase

(i).F

ora

sufficiently

high

tax(T

�I

o�

c),P

ersonO

ne’s

response

toA

ctivityTw

ois

no

longer

credible,and

Person

Two

will

switch

fromA

ctivityO

ne

toA

ctivityTw

o.T

his

policyben

efits

Person

Two

atth

eexpen

seof

Person

One.

Totalutility

chan

gesfrom

Vto

2V�

Is

�I

o ,apositive

chan

geif

Vexceeds

Is

�I

o . 29A

policyw

ithth

eopposite

effectis

atax

onA

ctivityTw

oitself.T

his

policyw

ould

benefi

tP

ersonO

ne

atth

eexpen

seof

Person

Two

incases

(ii)and

(iii).Inth

efirst

(second)case,a

taxin

excessofV

�I

s�

L(V

�I

s )in

duces

Person

Two

todesist

fromA

ctivityTw

o.T

his

policyw

ould

increase

totalutility,

inth

efirst

case,if

V�

28.A

noth

erbasis

forth

em

odelis

the

psychology

ofcogn

itivedisson

ance.

When

Person

Two

engages

inA

ctivityTw

o,she

challen

gesth

evalidity

ofP

ersonO

ne’s

beliefs,an

dP

ersonO

ne

suffersfrom

cognitive

dissonan

ce.To

remove

this

dissonan

ce,Person

One

may

actagain

stP

ersonTw

o.29.

Of

course,

such

a‘‘w

elfarean

alysis’’is

subject

toth

eusu

alcaveats

concern

ing

interperson

alcomparison

san

dth

em

easurability

ofutility.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

730

c�

Is

�L

,and,in

the

second

case,ifV

�I

o�

Is .F

inally,policies

may

chan

geth

eprescription

sth

emselves.A

rhetoricalcam

paign,

forexam

ple,m

aym

akeA

ctivityTw

om

oreloath

some

toG

reens,

leading

tohigh

ervalu

esofI

s and

Io an

dgreater

conform

ityto

the

prescriptions.

Of

course,

adifferen

tcam

paigncou

ldhave

the

oppositeeffect.

These

policiesare

identity

examples

ofth

econ

flict

ofth

eP

aretianL

iberal[S

en1970].

Itis

not

possibleto

protectP

ersonO

ne

against

the

externalities

caused

byP

ersonTw

o’sch

oiceof

Activity

Two

and

atth

esam

etim

eprotect

Person

Two

fromP

ersonO

ne’s

response.

There

isa

conflict

between

protecting

individu

alsw

ho

engage

incertain

activitiesan

dsu

ppressing

these

same

activitiesth

atm

aycau

seoth

ersdiscom

fortan

dan

xiety.

D.E

xtension

sto

the

Mod

eland

the

Defi

nition

of‘‘Situ

ations’’

Differen

tassu

mption

sabou

tiden

tity,pairings,an

din

forma-

tionallyield

potentially

interestin

gexten

sions

toth

em

odel.As

inth

ebasic

model,

individu

albeh

aviorw

ould

depend

onw

hat

sociologistsw

ould

callth

e‘‘situ

ation’’—

who

ism

atched

with

whom

and

inw

hat

context. 30

Inth

ebasic

model,everyon

esh

aredth

esam

eiden

tityan

dprescription

s,bu

tth

erecou

ldbe,

more

realistically,m

any

identities

amon

gth

epopu

lation.

Activities

One

and

Two

could

have

different

mean

ings

fordifferen

tpeople.

For

example,by

choosin

gA

ctivityTw

o,aperson

could

affirmher

identity

asa

Red.

People

could

alsoch

oose—m

oreor

lesscon

-sciou

sly—th

eiriden

titiesas

wellas

their

activities.These

choices

could

depend

onth

eprobability

ofdifferen

tm

atchin

gs,or

situa-

tions. 31

We

will

exploreth

ispossibility

belowin

our

study

ofpoverty

and

socialexclusion

.F

urth

ermore,pairin

gsneed

not

beexogen

ous,n

ortastes

and

prescriptions

know

n.In

fact,much

conflict

occurs

because

peoplew

ithdifferen

tprescription

sor

identities

come

into

contact.

Toavoid

conflict

and

lossesin

utility,people

may

wan

tto

match

with

those

who

share

the

same

identity

orfor

whom

actions

have

the

same

mean

ing.

Thus,

the

match

ing

processitself—

the

‘‘situa-

30.W

hen

anin

dividual’s

identity

isassociated

with

multiple

socialcatego-

ries,th

e‘‘situ

ation’’

could

determin

e,for

example,

which

categoriesare

most

salient.

31.C

hoice

could

alsodepen

don

frequen

cyof

certainaction

s.K

uran

[1998]con

siderseth

nically

symbolic

activitiesin

am

odelw

here

peoplecare

about

belongin

gto

aneth

nic

group.

When

greateroverall

resources

aredevoted

toan

ethnic

activity,an

individu

al’sm

arginal

utility

fromth

isactivity

canin

crease,leadin

gto

an‘‘eth

nifi

cation’’cascade.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y731

tions’’

inw

hich

agents

find

them

selves—can

been

dogenou

s,driven

byprescription

san

diden

tities.W

ew

illsee

this

outcom

ebelow

inou

rfirst,an

dperh

apsm

ostobviou

sapplication

.

IV.ID

EN

TIT

Y,G

EN

DE

R,A

ND

EC

ON

OM

ICS

INT

HE

WO

RK

PL

AC

E

An

identity

theory

ofgen

derin

the

workplace

expands

the

econom

ican

alysisof

occupation

alsegregation

.A

srecen

tlyas

1970,two-th

irdsof

the

United

States’fem

aleor

male

laborforce

wou

ldhave

had

tosw

itchjobs

toach

ieveoccu

pationalparity.T

his

measu

reof

occupation

alsegregation

remain

edvirtu

allyun-

chan

gedsin

ceth

ebegin

nin

gof

the

centu

ry.Yet,intw

enty

years,from

1970to

1990,this

figu

redeclin

edto

53percen

t. 32An

identity

modelpoin

tsto

chan

gesin

societalnotion

sofm

alean

dfem

aleas

am

ajorcau

se.T

he

modelw

epropose

captures

the

‘‘auras

ofgender’’[G

oldin1990a]

that

have

pervadedth

elabor

market.

Occu

pations

areassociated

with

the

socialcategories

‘‘man

’’an

d‘‘w

oman

,’’an

din

dividual

payoffsfrom

different

typesof

work

reflect

these

gender

associations.

This

model

canexplain

patterns

ofoccu

pa-tion

alsegregation

that

have

eluded

previous

models.

Italso

directlycaptu

resth

econ

sequen

cesofth

ew

omen

’sm

ovemen

tan

daffords

anew

econom

icin

terpretationofsex

discrimin

ationlaw

.Iden

tityalso

providesa

microfou

ndation

forearlier

models.

The

‘‘distaste’’ofm

enfor

workin

gw

ithw

omen

,asin

the

crudest

adaptations

ofracial

discrimin

ationm

odels[B

ecker1971;A

rrow1972],

canbe

understood

asdu

eto

lossin

male

identity

when

wom

enw

orkin

am

an’s

job.Sim

ilarlyw

omen

’sassu

med

lower

desirefor

laborforce

participation(as

inM

incer

and

Polach

ek[1974],B

ulow

and

Sum

mers

[1986],and

Lazear

and

Rosen

[1990])can

beunderstood

asth

eresu

ltofth

eiriden

tityas

hom

emakers. 33

A.T

he

Mod

el 34

There

aretw

osocial

categories,‘‘m

en’’

and

‘‘wom

en,’’

with

prescriptions

ofappropriate

activitiesfor

each.A

firm

wish

esto

32.See

Goldin

[1990a,C

hapter

3]for

historical

measu

resof

occupation

alsegregation

.See

Blau

,Sim

pson,

and

Anderson

[1998],w

ho

use

Cen

sus

Bureau

three-digit

classification

sofoccu

pations,for

1970–1990figu

res.33.

InB

ergman

n[1974],

male

employers

areaverse

tohirin

gw

omen

forparticu

larjobs

and

may

collude

tokeep

wom

enou

tof

high

paying

occupation

s,reservin

gth

egain

sfor

other

males.

Inou

rth

eory,th

esou

rceof

occupation

alsegregation

isem

piricallym

otivated—th

em

ainten

ance

ofgen

deriden

tityon

the

partofem

ployees.34.

An

appendix

with

complete

specification

ofthe

modelis

availablefrom

the

auth

orsupon

request.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

732

hire

laborto

performa

task.By

the

initialprescription

s,this

taskis

appropriateon

lyfor

men

;it

isa

‘‘man

’sjob.’’

Relative

toa

‘‘wom

an’s

job,’’wom

enlose

identity

inam

ount

Is

byperform

ing

such

work. 35

Inth

issitu

ation,m

alecow

orkerssuffer

aloss

Io . 36

They

may

relieveth

eiran

xietyby

taking

actionagain

stw

omen

coworkers, 37redu

cing

everyone’s

productivity.

Toavoid

these

productivity

losses,th

efirm

may

chan

gegen

der-jobassociation

sat

acost.

The

firm

islikely

tocreate

a‘‘w

oman

’sjob’’alon

gsideth

e‘‘m

an’s

job,’’rather

than

render

the

whole

taskgen

derneu

tral,w

hen

anew

jobdescription

canpiggyback

onexistin

gnotion

sofm

alean

dfem

ale. 38Aw

ell-know

nhistorical

example

illustrates.In

the

nin

eteenth

centu

ry,Horace

Man

n(as

Secretary

ofEdu

cationfor

Massach

usetts)tran

sformed

elemen

tarysch

oolteach

ing

into

aw

oman

’sjob,

arguin

gth

atw

omen

were

‘‘more

mild

and

gentle,’’

‘‘ofpu

rerm

orals,’’w

ith‘‘stron

gerparen

talim

pulses.’’ 39

Secon

darysch

oolteach

ing

and

schooladm

inistration

remain

edjobs

form

en.

The

model

alsoin

dicatesw

hy

gender-job

associations

may

persist.If

associations

aresectorw

ideor

econom

ywide,

and

not

firm

-specific,perfectly

competitive

firm

sw

illunderin

vestin

new

jobcategories.

Ben

efits

wou

ldaccru

eto

other

firm

s.In

the

absence

ofmarket

power

ortech

nologicalch

ange,a

shift

insocial

attitudes

and

legalinterven

tionw

ould

benecessary

forch

anges

inem

ploymen

tpattern

s.T

he

modeleasily

extends

toth

edecision

toparticipate

inth

elabor

force.If

wom

en’s

identity

isen

han

cedby

work

inside

the

hom

e,th

eyw

illhave

lower

laborforce

attachm

ent

than

men

.H

istorically,fem

alelabor

forceparticipation

rates,relative

tom

alerates,h

avebeen

bothlow

eran

dm

orecyclically

variable.

B.Im

plications

forL

aborM

arketOu

tcomes

This

identity

model

explains

employm

ent

patterns

arising

fromassociation

sbetw

eengen

deran

dtype

ofw

ork.T

hese

pat-

35.B

lauan

dF

erber[1986,

Chapter

7]also

discuss

the

‘‘psychic

costs’’in

curred

bya

wom

an(m

an)

workin

gin

ajob

requirin

g‘‘m

asculin

e’’(‘‘femin

ine’’)

traits.36.G

oldin[1990b]

considers

am

odelw

here

men

lose‘‘statu

s’’when

wom

enw

orkon

their

jobsbecau

seth

ejobs

arerevealed

tobe

lessdifficu

ltor

physically

deman

ding.

37.W

ehave

alreadyseen

such

emotion

san

dbeh

aviorin

Pierce’s

[1995]law

firm

and

Padavic’s

[1991]power

plant.S

chultz

[1998]relatesa

plethora

ofsimilar

cases.38.A

firm

with

market

power

will

earna

furth

erbon

us

fromoccu

pational

segregationin

the

formofw

agediscrim

ination

.39.

See

quotation

ofMan

nin

Sugg

[1978,p.74],and

other

An

nu

alReports

byM

ann.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y733

terns

gobeyon

dw

hat

canbe

explained

byw

omen

’sassu

med

lower

laborforce

attachm

ent

asin

Min

ceran

dP

olachek

[1974],where

wom

enw

orkin

occupation

sth

atrequ

irelittle

investm

ent

infirm

-specific

hum

ancapital. 40

Inou

rm

odel,wom

enw

illdomin

atejobs

whose

requirem

ents

match

constru

edfem

aleattribu

tesan

din

feriorsocialstatu

s;men

eschew

them

.Historically,th

reeoccu

pations

illustrate:secretar-

ies(97.8

percent

female

in1970

41)have

oftenbeen

called‘‘office

wives,’’an

delem

ents

ofsexu

alityare

inscribed

inth

ew

orking

relationsh

ip(boss

�m

ale,secretary�

female)[M

acKin

non

1979;P

ringle

1988].Secretaries

areexpected

toserve

their

bosses,with

deference,

and

tobe

attentive

toth

eirperson

alneeds

[Davies

1982;K

anter

1977;P

ierce1996].

Elem

entary

school

teachers

(83.9percen

tfem

ale),in

contrast

tosecon

darysch

oolteach

ers(49.6

percent

female),

aresu

pposedto

carefor

young

children

.N

urses

(97.3percen

tfem

ale)are

supposed

tobe

tender

and

carefor

patients,

asw

ellas

bedeferen

tialto

doctors[F

isher

1995;W

illiams

1989].In

our

model,w

omen

donot

enter

male

professions

because

ofgen

derassociation

s.H

istorically,m

any

male

professions

have

required

similar

levelsof

education

and

trainin

gto

female

professions

and

could

have

beenam

enable

topart-tim

ean

din

termitten

tw

ork.C

ontrast

nursin

gan

dteach

ing

with

account-

ing

and

law.

All

require

collegedegrees

and

certification

,an

dsom

etimes

have

tenure

and

experience-based

pay.Only

the

verytop

ofth

eseprofession

shave

required

contin

uity

inem

ploymen

tan

dfu

ll-time

work.

Rhetoric

surrou

ndin

gjob

shifts

fromm

aleto

female

furth

erdem

onstrates

the

salience

ofgender-job

associations.T

he

recruit-

men

tofwom

enin

to‘‘m

en’s

jobs’’durin

gW

orldW

arII,for

example,

was

accompan

iedby

officialpropagan

daan

dpopu

larliteratu

repictu

ring

wom

entakin

gon

factoryw

orkw

ithou

tlossoffem

inin

ity[M

ilkman

1987;Hon

ey1984;P

ierson1986].In

addition,th

ejobs

40.T

he

empirical

evidence

forth

ishum

ancapital

explanation

ism

ixed(see

Blau

,Sim

pson,

and

Anderson

[1998]for

review).

Oth

erth

eoriesbased

onlow

workplace

attachm

ent

ofw

omen

inclu

deL

azearan

dR

osen[1990],w

here

occupa-

tional

segregationis

aform

ofstatistical

discrimin

ation;

workers

inth

em

aleoccu

pations,i.e.,w

ithhigh

laborforce

attachm

ent,are

targetsfor

promotion

,and

those

inth

efem

aleoccu

pations

arenot.In

Bulow

and

Sum

mers

[1986],primary-

sectorfirm

sm

ust

payw

omen

high

erw

ageprem

ium

sto

prevent

them

fromsh

irking

because

wom

enare

more

likelyto

quit

their

jobs.These

firm

s,therefore,

preferhirin

gm

ento

wom

en.

41.See

Blau

,Sim

pson,

and

Anderson

[1998,A

ppendix

A-1]

forth

esean

dfollow

ing

figu

res.Allfi

gures

here

arefor

1970.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

734

were

portrayedas

temporary;

only

the

wartim

eem

ergency

ex-cu

sedth

eviolation

ofthe

usu

algender

prescriptions.

C.E

ffectsofth

eW

omen

’sM

ovemen

t

The

model

givesa

theoretical

structu

refor

how

the

wom

en’s

movem

entm

ayhave

impacted

the

laborm

arket.The

movem

ent’s

goalsin

cluded

reshapin

gsocietal

notion

sof

femin

inity

(and

mascu

linity)

and

removin

ggen

derassociation

sfrom

tasks,both

inth

ehom

ean

din

the

workplace.

Inth

em

odel,su

chch

anges

wou

lddecrease

wom

en’s

gains

(men

’slosses)

iniden

tityfrom

hom

emakin

g,an

ddecrease

the

identity

lossI

sof

wom

en(m

en)

workin

gin

traditionally

men

’s(w

omen

’s)jobs,

asw

ellas

the

accompan

ying

externalities

Io .T

hese

shifts

wou

ldin

creasew

om-

en’s

laborforce

participationan

dlead

toa

convergen

ceof

male

and

female

jobten

ure

rates.M

orew

omen

(men

)w

ould

work

inpreviou

slym

ale(fem

ale)jobs.A

llthese

outcom

esare

observedcoin

cidentalw

ithan

dfollow

-in

gth

ew

omen

’sm

ovemen

t. 42Gen

der-jobassociation

sdim

inish

ed,refl

ectedin

chan

gesin

langu

age(e.g.,

firem

enbecam

efirefi

ght-

ers).In

1998th

em

edianjob

tenure

ofem

ployedw

omen

over25

was

0.4years

lower

than

that

ofm

en;in

1968th

atgap

had

been3.3

years. 43C

han

gesin

sexcom

positionw

ithin

occupation

sac-

counted

forth

em

ajorsh

areofdeclin

ein

occupation

alsegregationfrom

1970–1990[B

lau,S

impson

,and

Anderson

,1998].Ofth

e45

three-digit

Cen

sus

occupation

sth

atw

ere0.0

percent

female

in1970,

only

one

(supervisors:

brickmason

s,ston

emason

s,an

dtile

setters)was

lessth

an1

percentfem

aletw

enty

yearslater. 44M

any

incu

rsions

offem

alesin

tom

ale-domin

atedprofession

sw

erevery

large.Con

sideragain

accountin

gan

dlaw

.In1970

(1990)females

were

24.6(52.7)

percent

ofau

ditorsan

daccou

ntan

ts,an

d4.5

(24.5)percentoflaw

yers.Not

only

didth

eproportion

ofwom

enin

men

’sjobs

increase,bu

tso

didth

eproportion

ofm

enin

wom

en’s

jobs(albeit

much

lessdram

atically). 45O

fth

etriu

mvirate

of

42.T

he

Fem

inin

eM

ystique

was

publish

edin

1963,and

the

Nation

alOrgan

i-zation

forW

omen

was

founded

in1966.

43.3.8

yearsfor

men

versus

3.4for

wom

enin

1998[U

nited

States

Depart-

men

tof

Labor,

1998];7.1

yearsfor

men

versus

3.8for

wom

enin

1968.[S

ource:

calculation

fromTable

A,U

.S.D

epartmen

tof

Labor,S

pecialL

aborF

orceR

eport112,J

obTen

ure

ofW

orkers,Jan

uary

1968.]T

he

figu

resfor

the

two

yearsare

not

strictlycom

parable;in

1968th

equ

estionasked

forth

etim

eelapsed

since

the

beginnin

gof

the

curren

tjob,in

1998sin

ceth

ecu

rrent

employer.M

edianm

alejob

tenure

has

alsobeen

considerably

affectedby

shifts

inth

eage

distribution

ofth

ew

orkforce,bothbecau

seofdem

ographic

shifts

and

alsoearly

retiremen

t.44.

Sou

rce:Blau

,Sim

pson,an

dA

nderson

[1998,Appen

dixA

-1].45.

See

Blau

,Sim

pson,an

dA

nderson

[1998,Table3

and

Appen

dixA

-1].

EC

ON

OM

ICS

AN

DID

EN

TIT

Y735

explanation

sfor

such

increases—

technology,

endow

men

ts,an

dtastes—

elimin

ationm

akestastes

the

leading

suspect,sin

ceth

erew

asno

dramatic

chan

gein

technology

oren

dowm

ents

that

wou

ldhave

caused

such

increased

mixin

gon

the

job. 46L

egalin

itiativesdiscu

ssednext

reflect

such

chan

gesin

tastes.

D.G

end

er-Job

Association

san

dS

exD

iscrimin

ationL

aw

Legal

interpretation

sof

sexdiscrim

ination

correspond

toearlier

econom

icm

odelsas

well

asou

row

n.T

itleV

IIof

the

Civil

Righ

tsA

ctof1964

makes

itunlaw

fulfor

anem

ployerto

discrimi-

nate

‘‘against

any

individu

al...with

respectto

...compen

sation,

terms,

condition

s...

ofem

ploymen

t’’or

‘‘to[adversely]

limit,

segregate,orclassify

his

employees

...because

of...sex.’’ 47Atits

most

basic,this

lawproh

ibitsa

discrimin

atoryexercise

of‘‘tastes’’again

stw

omen

(analogou

sto

Becker

[1971]an

dA

rrow[1972]).

Cou

rtsalso

interpretT

itleV

IIas

outlaw

ing

statisticaldiscrimin

a-tion

bysex

orcriteria

correlatedw

ithsex,

evenw

hen

wom

enon

averagelack

adesirable

jobqu

alification

.D

iscrimin

atoryhirin

gbecau

seof

wom

en’s

presum

edlow

erw

orkplaceattach

men

t,asin

Lazear

and

Rosen

[1990],w

asprecisely

the

issue

addressedin

Ph

illipsv.M

artin-M

arietta. 48O

ur

model,

where

sexdiscrim

ination

occurs

because

jobshave

gender

associations,correspon

dsto

aw

iderin

terpretationof

Title

VII.

This

interpretation

isat

the

forefront

ofcu

rrent

legaldebate

and

issu

pportedby

anum

berofpreceden

ts.InD

iazv.P

anA

merican

World

Airw

ays, 49the

Cou

rtoutlaw

edsex

bans

inhirin

g.T

he

airline

originally

pleadedfor

their

prohibition

ofm

alefligh

tatten

dants

because

wom

enw

erebetter

at‘‘th

enon

mech

anical

aspectsofth

ejob.’’B

utth

isassociation

ofgender

with

the

jobw

asdisallow

edon

appealsince

femin

ine

traitsw

eredeem

edirrelevan

tto

the

‘‘primary

function

orservices

offered’’(citedin

MacK

innon

[1979,p.180]).Price

Waterh

ouse

v.Hopkin

s50

seta

precedent

forw

orkersalready

hired.

The

plaintiff

had

beenden

ieda

partner-

ship

afternegative

evaluation

sfor

her

mascu

line

deportmen

t.T

he

Suprem

eC

ourt

ruled

that

‘‘anem

ployerw

ho

objectsto

46.C

ompu

tersare

used

inten

sivelyin

fewof

the

occupation

sw

ithm

ajorch

anges

inm

ix.47.

42U

.S.C

.§§2000e–2000e17

[1982],Section

s703(a)(1)an

d703(a)(2).

48.442

F.2d385

(5thC

ir.1971),cert.den

ied,404

U.S

.950(1971).G

riggsv.

Du

keP

ower,

401U

.S.

424(1971),

arace

discrimin

ationcase,

isan

importan

tpreceden

tou

tlawin

gtest

results

and

other

criteriacorrelated

with

raceor

gender

asem

ploymen

tscreen

s.49.

442F.2d

385(5th

Cir.)cert.d

enied

,404U

.S.950

(1971).50.

490U

.S.228

(1989).

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

736

aggressiveness

inw

omen

but

whose

positions

require

this

traitplaces

wom

enin

anin

tolerablean

dim

permissible

Catch

22’’(cited

inW

urzbu

rgan

dK

lonoff

[1997,p.

182]).C

aseshave

alsoin

volvedharassm

ent

ofw

omen

workin

gin

men

’sjobs

as,in

the

termin

ologyof

our

model,

male

coworkers

protectth

emselves

fromloss

ofidentity

Io .B

erkman

v.City

ofNew

York51rein

stateda

firefi

ghter

who

had

beendism

issedbecau

seof

substan

dardw

orkperform

ance.

The

Cou

rtru

ledth

atth

ein

terference

and

harass-

men

tby

her

male

coworkers

made

itim

possiblefor

her

toperform

her

jobadequ

ately[S

chultz

1998,p.1770].This

expansive

inter-

pretationof

a‘‘h

ostilew

orken

vironm

ent,’’a

categoryof

sexual

harassm

entw

hich

isin

turn

acategory

ofsexdiscrim

ination

,has

beenexception

al.Judges

have

viewed

sexual

desireas

anessen

-tial

elemen

tof

sexual

harassm

ent.

How

ever,Sch

ultz

[1998]an

dF

ranke

[1995]argu

eth

atan

yharassm

ent

derivedfrom

gender

prescriptions

has

discrimin

atoryim

plications

(asdepicted

inou

rm

odel)and

areth

us

violations

ofTitle

VII.

V.ID

EN

TIT

YA

ND

TH

EE

CO

NO

MIC

SO

FE

XC

LU

SIO

NA

ND

PO

VE

RT

Y

This

sectionw

illcon

sideriden

tityan

dbeh

aviorin

pooran

dsocially

excluded

comm

unities.

Inan

adaptationof

the

previous

model

ofG

reens

and

Reds,

peoplebelon

ging

topoor,

sociallyexclu

dedgrou

psw

illch

ooseth

eiriden

tity.G

reens

identify

with

the

domin

ant

cultu

re,while

those

with

Red

identity

rejectit

and

the

subordin

ateposition

assigned

toth

oseofth

eir‘‘race,’’class,or

ethnicity. 52

From

the

point

ofview

ofth

osew

ithG

reeniden

tities,R

edsare

oftenm

aking

badecon

omic

decisions;th

eym

ighteven

bedescribed

asen

gaging

inself-destru

ctivebeh

avior.Taking

drugs,

joinin

ga

gang,an

dbecom

ing

pregnan

tatayou

ng

ageare

possiblesign

sof

aR

ediden

tity.T

his

aspectof

behavior

has

not

beenexplored

inpreviou

sm

odels,but

itis

implicit

inW

ilson’s

account

ofblack

ghetto

poverty[1987,

1996].It

alsois

implicit

inevery

51.580

F.Supp.

226(E

.D.N

.Y.1983),

aff’d,

755F.

2d913

(2dC

ir.1985).

Berkm

anfollow

edth

eexpan

siveview

inM

cKin

ney

v.D

ole,765

F.2d

1129(D

.C.

Cir.1985),th

at‘‘an

yharassm

entor

unequ

altreatmen

tofan

employee

orgrou

pof

employees

thatw

ould

notoccu

rbu

tforth

esex

ofthe

employee

orem

ployeesm

ay,ifsufficien

tlypattern

edor

pervasive,com

prisean

illegalcon

ditionof

employm

ent

under

Title

VII’’(cited

inSch

ultz

[1998,p.1733]).52.

Much

literature

oniden

tityan

dsocial

exclusion

argues

that

domin

ant

groups

define

them

selvesvis-a-vis

‘‘other(s),’’

and

mem

bersof

the

domin

ant

(excluded)

groups

benefi

t(lose)—

materially

and

psychologically—

fromth

ediffer-

entiation

.F

ordiscu

ssionof

different

approaches

toth

estu

dyof

socialdifferen

cean

dracism

,seeW

etherell[1996].

EC

ON

OM

ICS

AN

DID

EN

TIT

Y737

study

that

finds

signifi

cant

dum

my

variablesfor

‘‘race,’’after

adjustm

ent

foroth

erm

easures

ofsocioecon

omic

status.

The

Green

/Red

model

ofth

issection

offersan

explanation

forth

esign

ifican

ceof

such

dum

my

variables.F

urth

ermore,

ityields

aless

mon

olithic

viewof

povertyth

ancu

rrent

econom

icth

eoriesth

atem

phasize

conform

ity(e.g.,

Akerlof

[1997]an

dB

rockan

dD

urlau

f[1995]).

A.M

otivationfor

Mod

el

Our

modelrefl

ectsth

em

any

ethnograph

icaccou

nts

of‘‘oppo-sition

al’’identities

inpoor

neigh

borhoods.M

acLeod’s

[1987]study

ofteen

agersin

aB

ostonarea

hou

sing

project,for

example,

contrasts

the

murderou

san

dalcoh

olicH

allway

Han

gersto

their

obedient

and

athletic

peers,the

Broth

ers.InL

earnin

gto

Labou

rW

illis[1977]describes

the

antagon

ismbetw

eenth

eunru

ly‘‘lads’’

and

the

dutifu

l‘‘earh

oles’’ina

workin

g-classE

nglish

secondary

school.

Sim

ilarly,W

hyte’s

[1943]description

ofB

oston’s

ItalianN

orthE

nd

circa1940

contrasts

the

Corn

erB

oysto

the

College

Boys.Yet

earlier,turn

-of-the

centu

ryaccou

nts

ofth

eIrish

inth

eU

nited

States

contrast

the

‘‘lacecu

rtain’’Irish

ofpoor

districtsto

their

neigh

bors(see,e.g.,M

iller[1985]).

Our

model

furth

erevokes

the

psychological

effectsof

socialexclu

sionin

the

colonial

experience

analyzed

byB

habh

a[1983]

and

Fan

on[1967],an

din

the

context

ofAfrican

-Am

ericans

inth

eU

nited

States

byA

nderson

[1990],Baldw

in[1962],C

lark[1965],

DuB

ois[1965],F

razier[1957],H

annerz

[1969],Rain

water

[1970],W

ilson[1987,

1996],an

doth

ers.In

these

settings,

individu

alsfrom

particular

groups

cannever

fully

fit

the

idealtype,the

ideal‘‘G

reen,’’of

the

domin

ant

cultu

re.Som

ein

excluded

groups

may

tryto

‘‘pass’’orin

tegratew

ithth

edom

inan

tgrou

p,butth

eydo

sow

itham

bivalence

and

limited

success. 53

Aseries

ofau

tobiogra-ph

iestells

ofth

epain

and

anger

ofdiscoverin

gth

aton

eis

not

really‘‘G

reen.’’F

ormer

New

YorkT

imes

editorM

elWatkin

s[1998]

titlesth

ech

apteron

his

freshm

anyear

atC

olgateas

‘‘stranger

ina

strange

land.’’G

andh

i[1966],

Fan

on[1967],

Fulw

ood[1996],

Staples

[1994],and

Rodrigu

ez[1982]

allrelate

strikingly

similar

experiences

ofperceived

orreal

rejectionan

dalien

ation.

This

socialexclu

sionm

aycreate

acon

flict:

how

tow

orkw

ithin

the

domin

antcu

lture

with

outbetrayin

gon

eself.As

JillNelson

[1993,

53.In

deed,th

ew

ordpassin

gitself

ispejorative

and

evokesa

penum

braof

reactions

tobein

goth

erth

anon

e’s‘‘tru

e’’self.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

738

p.10]explains

her

exhau

stionafter

alon

gday

ofinterview

ing

fora

jobat

Th

eW

ashin

gtonP

ost:

I’vealso

beendoin

gth

estan

dardN

egrobalan

cing

actw

hen

itcom

esto

dealing

with

white

folks,which

involves

sufficiently

blurrin

gth

eedges

ofmy

being

soth

atth

eydon

’tfeelin

timidated,w

hile

simultan

eously

holdin

gon

tom

yin

tegrity.There

isa

thin

line

between

Uncle-Tom

min

gan

dM

au-M

auin

g.To

falloff

that

line

canm

eandisaster.

On

one

sidelies

employm

ent

and

self-hatred;

onth

eoth

er,th

eequ

allydu

bious

hon

orof

unem

ploymen

tw

ithin

tegrity.

These

reactions,itm

ustbe

emph

asized,reflecth

owdom

inan

tgrou

psdefi

ne

them

selvesby

the

exclusion

ofothers.T

he

creationan

devolu

tionof

such

socialdifferen

cesare

the

subject

ofm

uch

historical

research.Said

[1978]docu

men

tsth

eem

ergence

ofth

eW

esternidea

ofth

e‘‘O

riental,’’

acon

ceptth

athad

signifi

cant

implication

sfor

colonialism

.Inth

eU

nited

States

Roediger

[1991]an

doth

erhistorian

ssh

owhow

workers

ofE

uropean

descent

inth

enin

eteenth

centu

ryin

creasingly

were

defined

as‘‘w

hite.’’P

riorto

Em

ancipation

,this

identity

evokedth

econ

trastbetw

eenw

hite

freedoman

dA

frican-A

merican

enslavem

ent.

Inth

em

odelw

econ

struct,

the

keyin

teractionis

between

such

socialdifferen

cesan

dth

eadoption

ofopposition

aliden

titiesby

those

inexclu

dedgrou

ps.L

ackof

econom

icopportu

nity

may

alsocon

tribute

toth

ech

oiceofan

oppositionaliden

tity.Wilson

[1987,1996]underscores

the

relationbetw

eenth

edeclin

ein

remunerative

unskilled

jobs,th

eloss

ofself-respect

bym

enw

ho

cannot

support

their

families,

and

the

risein

inner

citycrim

ean

ddru

gabu

se.T

his

processis

illustrated

inm

icrocosmby

‘‘Rich

ard’’inT

ally’sC

orner

[Liebow

1967].U

nable

tofind

decent-payin

gw

ork,he

abandon

edhis

family

and

joined

Tally’sgrou

pof

idlerson

the

streetcorn

er.B

yadoptin

ga

different

identity,R

ichard

no

longer

sufferedth

egu

iltofa

failedprovider. 54

Red

activitieshave

negative

pecuniary

externalities.

Rich

-ard’s

wife

and

children

had

tofind

alternative

mean

sof

support.

The

prime

goalofthe

‘‘lads’’inW

illis’secondary

schoolw

asto

geta‘‘laff,’’th

rough

vandalism

,picking

figh

ts,and

return

ing

drunk

tosch

oolfrom

the

localpu

b.Runnin

ga

school

with

ladsis

difficult.

The

situation

corresponds

toth

eextern

alitiesin

Ben

abou’s

[1993,1996]

models

ofhigh

schoolin

gcosts

inpoor

neigh

borhoods.

54.See

Mon

tgomery

[1994]for

anin

terpretationof

Rich

ard’sbeh

aviorin

terms

ofcognitive

dissonan

ce.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y739

Furth

erextern

alitiesaccru

efrom

drug

dealing,crim

e,and

other

‘‘pathological’’beh

avior.In

our

model,

there

arealso

identity-based

externalities.

AR

edis

angered

bya

Green

’scom

plicityw

ithth

edom

inan

tcu

lture,

while

aG

reenis

angered

bya

Red’s

‘‘breaking

the

rules.’’A

gaincon

siderW

illis’ladsan

dearh

oles.As

the

ladsdefi

ne

them

selvesin

contrast

toth

eearh

oles,th

eearh

olesdefi

ne

them

selvesin

con-

trastto

the

lads.T

he

earholes

areeven

more

proestablishm

ent

than

the

teachers—

feeling

that

the

teachers

shou

ldbe

stricter.T

he

lads,in

turn

,bait

the

earholes.

This

situation

isju

ston

e(relatively

tame)

example

ofhow

interaction

between

the

two

groups

generates

antagon

ismon

bothsides.

B.Id

entity

Mod

elofPoverty

and

SocialE

xclusion

As

inth

eprototype

model,th

ereare

two

activities,One

and

Two.A

ctivityO

ne

canbe

thou

ghtofas

‘‘workin

g’’and

Activity

Two

as‘‘n

otw

orking.’’T

here

isa

largecom

munity,n

ormalized

tosize

one,

ofin

dividuals.

The

econom

icretu

rnto

Activity

One

forin

dividual

iis

vi

which

we

assum

eis

uniform

lydistribu

tedbetw

eenzero

and

one,

torefl

ectheterogen

eityin

the

population

and

toen

sure

interior

solution

s.The

econom

icretu

rnto

Activity

Two

isnorm

alizedto

zero.A

sfor

identity,

there

aretw

osocial

categories,G

reenan

dR

ed.AG

reensuffers

aloss

iniden

tityr,represen

ting

the

extentto

which

someon

efrom

this

comm

unity

isnot

acceptedby

the

domin

ant

group

insociety.

Those

with

the

lessadaptive

Red

identity

donot

sufferth

isloss.B

ehavioral

prescriptions

sayth

atG

reens

(Reds)sh

ould

engage

inA

ctivityO

ne

(Two).T

hus,a

Green

(Red)

losesiden

tityfrom

Activity

Two

(One)

inam

ount

Is G(I

s R). 55B

ecause

Reds

rejectth

edom

inan

tG

reencu

lture,

they

arealso

likelyto

have

lower

econom

icretu

rns

toA

ctivityO

ne

than

Green

s. 56A

Red

individu

ali

will

only

earnv

i�

afrom

Activity

One,

asw

ellas

sufferth

eloss

Is R.

There

arealso

identity

externalities

when

Green

san

dR

edsm

eet.AG

reen(R

ed)suffersa

lossI

o G(Io R).

Inaddition

,R

edsw

ho

have

chosen

Activity

Two

55.W

ediscu

ssbelow

the

possibilityof

aR

ediden

tityw

here

individu

alscan

bothreject

the

domin

antcu

lture

and

atth

esam

etim

edo

not

loseI

s Rfrom

Activity

One.56.

Wilson

[1996,C

hapter

5]docu

men

tsth

edifficu

ltiesth

atem

ployersperceive

inhirin

gem

ployeesfrom

the

inner

city.F

romth

evan

tagepoin

tof

our

model,

itdoes

not

matter

wheth

erth

eperceived

problems,

parameterized

bya,

reflectrealdifferen

cesin

productivity

orth

oseth

atarem

erelyim

agined

because

ofth

em

ismatch

ofthe

employers’an

dth

eem

ployees’attitudes.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

740

impose

apecu

niary

externality

kon

those

who

have

chosen

Activity

One.

Each

personi

chooses

aniden

tityan

dactivity,

giventh

ech

oicesofeveryon

eelse

inth

ecom

munity.W

eassu

me

that

peoplecan

not

modify

their

identity

oractivity

foreach

individu

alen

counter.

Rath

er,in

dividuals

choose

aniden

tityan

dactivity

tom

aximize

expectedpayoffs,given

the

probabilitiesof

encou

nters

with

Green

sw

ho

choose

Activity

One,

Green

sw

ho

choose

Two,

Reds

who

choose

One,an

dR

edsw

ho

choose

Two.

C.E

quilibria

and

Interpretation

Equ

ilibriaof

this

model

show

how

socialin

teractionw

ithin

the

comm

unity

and

socialexclu

sionfrom

the

domin

ant

group

determin

eth

eprevalen

ceof

Red

identities

and

Activity

Two

behavior. 57

An

All-G

reenE

quilibriu

m(everyon

eis

Green

and

engages

inA

ctivityO

ne)

exists,if

and

only

ifth

eloss

inG

reeniden

tity,r,

fromexclu

sionfrom

the

domin

ant

group

issm

allerth

anth

edifficu

ltyof

being

Red

ina

comm

unity

ofG

reens,

Io R.

Figu

reII

show

sth

iscon

ditionin

the

areaabove

the

45°lin

efrom

the

origin.F

orhigh

erlevels

ofr,equ

ilibriam

ust

involve

some

inth

ecom

munity

adopting

aR

ediden

tity.T

he

non

existence

ofth

eA

ll-Green

equilibriu

mreveals

adifferen

cein

the

predictions

ofth

ism

odelan

dpreviou

sm

odelsof

behavior

inpoor

neigh

bor-hoods.H

ere,socialexclu

sion(r

�0)

will

leadsom

epeople

inth

ecom

munity

toadopt

anopposition

aliden

tityan

dA

ctivityTw

obeh

avior,evenin

the

absence

ofcon

formity-gen

erating

externali-

ties(i.e.,I

o R�

Io G

�k

�0).

Ina

Mixed

Equ

ilibrium

ofour

model,som

ein

the

comm

unity

choose

Activity

One

and

aG

reeniden

tity,bu

toth

ersch

ooseA

ctivityTw

oan

dR

ediden

tities.T

his

equilibriu

marises

forin

termediate

levelsof

r(in

the

areabetw

eenth

etw

oupw

ard-slopin

glin

esin

Figu

reII).

The

equilibriu

madoption

ofR

ediden

titiesan

dA

ctivityTw

obeh

aviorcaptu

resth

eself-destru

ctivebeh

aviorof

the

underclass

central

tosociological

study,

but

contrary

tostan

dardecon

omic

thin

king.R

ainw

ater[1970,p.3]

sum

marized

his

classicstu

dyof

ghetto

poverty:‘‘w

hite

cupidity

createsstru

ctural

condition

shigh

lyin

imicalto

basicsocialadaptation

tow

hich

Negroes

adapt

57.F

ullan

alysisof

the

modelis

availablefrom

the

auth

orsupon

request.In

the

analysis

we

make

the

simplifyin

gassu

mption

that

Is G

�k

soth

atan

y-on

ew

ho

chooses

aG

reeniden

tityw

illch

ooseA

ctivityO

ne.

We

alsoassu

me

thatallparam

etersare

strictlypositive

andless

thanunity

andthatI

s R�

a�

k�

1.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y741

bysocial

and

personal

responses

which

serveto

sustain

the

individu

alinhis

punish

ing

world

butalso

togen

erateaggressive-

ness

toward

the

selfand

others

which

results

insufferin

gdirectly

inflicted

byN

egroeson

them

selvesan

don

others.’’W

hile

Activity

One

ism

aximizin

gto

someon

ew

itha

Green

identity,

itis

not

maxim

izing

tosom

eone

with

aR

ediden

tity.The

‘‘self-destructive’’

Red

behavior

isnot

the

result

ofin

dividual

‘‘irrationality,’’bu

tin

steadderives

fromlow

econom

icen

dowm

ents

and

ahigh

degreeofsocialexclu

sion.

Com

parativestatics

ofth

em

ixedequ

ilibrium

captures

Wil-

son’s

[1987,1996]an

alysisof

ghetto

poverty.An

out-m

igrationof

the

middle

class(th

osew

ithhigh

return

sv

iin

the

model)

will

result

infu

rther

adoptionof

Red

identities

amon

gth

erem

ainin

gpopu

lation.A

lso,when

work

disappears,there

willbe

adow

nw

ard

FIG

UR

EII

Equ

ilibriain

ModelofP

overtyan

dSocialE

xclusion

This

figu

resh

ows

ranges

ofparam

etervalu

esfor

three

different

equilibria:

All-G

reenw

here

everyone

isG

reenan

dch

oosesA

ctivityO

ne;

Mixed

where

Green

sch

ooseA

ctivityO

ne

and

Reds

choose

Activity

Two;

All-R

edw

here

everyone

isR

edan

dsom

ech

ooseA

ctivityO

ne

and

others

choose

Activity

Two.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

742

shift

indistribu

tionof

payoffsfrom

Activity

One.

This

shift

will

alsoin

creaseth

ein

cidence

ofActivity

Two

and

Red

identities.

Inan

All-R

edE

quilibriu

m,som

ein

dividuals

choose

Activity

One

and

conform

with

the

domin

ant

group

interm

sof

econom

icbeh

avior,butallch

oosean

oppositionalR

ediden

tity.This

equilib-

rium

arisesw

hen

ahigh

lossfrom

being

Green

inan

all-Red

comm

unity,I

o G,complem

ents

high

levelsof

socialexclu

sion,r

(inth

earea

toth

erigh

tof

the

verticallin

ein

Figu

reII). 58

This

equilibriu

mis

alsoach

ievedw

itha

lowvalu

eof

Is R

and,

thus,

providesan

interpretation

ofsocial

movem

ents

that

may

arisefrom

exclusion

.Som

eseparatist

leaders,such

asM

alcolmX

and

Lou

isF

arrakhan

,have

advanced

anopposition

alRed

identity

but

atth

esam

etim

ehave

triedto

chan

geassociated

prescriptions,

resultin

gin

alow

erI

s R.Inth

esem

ovemen

ts,Activity

One

doesnot

imply

complicity

with

the

domin

ant

group.R

ather,self-restrain

t,edu

cation,an

dem

ploymen

tare

am

eans

forin

dividual

advance-

men

tan

dcom

munity

liberation.

D.F

urth

erL

essons

fromth

eM

odel

The

model

and

itssolu

tionalso

affordin

terpretations

ofpolicies

designed

toredu

cepoverty

and

the

effectsof

socialexclu

sion.

First,

the

model

indicates

why

residential

JobC

orpspro-

grams

may

succeed

while

other

trainin

gprogram

sfail

[Stan

ley,K

atz,and

Kru

eger1998].A

ccording

toth

em

odel,taking

trainees

outofth

eirneigh

borhoods

wou

ldelim

inate,atleastfor

atim

e,the

negative

effectsof

interaction

with

those

with

Red

identities.

Moreover,

being

ina

different

locationm

ayredu

cea

trainee’s

directloss

rfrom

being

Green

and

pursu

ing

Activity

One.T

hat

is,th

isloss

may

beboth

individu

al-specific

and

situation

al,an

dleavin

ga

poorneigh

borhood

islikely

togen

eratea

lower

rth

anoth

erwise.In

asom

ewhatcon

trolledexperim

ent,th

eU

.S.govern

-m

enttried

tosave

mon

eyw

ithJO

BSTA

RT,w

hich

preservedm

any

ofth

efeatu

resof

JobC

orpsexcept

the

expensive

hou

sing

oftrain

ees.F

ollow-u

pstu

diesof

JOB

STA

RT

show

littleor

no

improvem

entin

employm

entor

earnin

gs. 59

58.It

overlapsth

eregion

sof

other

equilibria

because

this

condition

isin

dependen

tof

Io R,u

nlike

those

forth

eabove

equilibria

where

aR

edw

ould

sufferth

eloss

fromin

teracting

with

Green

s.59.

The

Cen

terfor

Em

ploymen

tan

dTrain

ing

inSan

Josew

asth

eon

erem

arkableexception

.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y743

Secon

d,th

em

odelaffords

anin

terpretationof

different

education

initiatives

form

inority

studen

ts.L

ikeJob

Corps,

the

Cen

tralP

arkE

astSecon

darySch

ool(C

PE

SS)

inE

astH

arlemm

aysu

cceedbecau

seit

separatesG

reenstu

dents

fromR

edstu

dents.S

tuden

ts,forexam

ple,mustapply

toth

esch

ool,indicat-

ing

their

and

their

parents’

willin

gness

toadopt

itsru

les(see

Fliegel

[1993]an

dM

eier[1995]

forth

isan

doth

erdetails).

Anoth

erin

terpretationofC

PE

SS

and

other

successes

(e.g.,Com

er[1980]in

New

Haven

)parallelsth

elogic

ofthe

all-Red

equilibriu

mw

here

some

peoplenon

etheless

pursu

eA

ctivityO

ne.T

he

schools

takem

easures

toredu

ceth

eloss

iniden

tityofR

edstu

dents,I

s R,inactivities

such

aslearn

ing

Stan

dardE

nglish

. 60D

elpit’s[1995]

award-w

innin

gbook

Oth

erP

eople’sC

hild

renproposes

num

erous

ways

toredu

ceth

ealien

ationth

atm

inority

studen

tsm

ayexperi-

ence

insch

ool.F

inally,

the

model

illum

inates

aset

ofissu

esin

the

affirma-

tiveaction

debate.M

uch

ofth

isdebate

concern

sth

esu

ccessor

failure

ofspecifi

cprogram

s(see,

e.g.,D

ickens

and

Kan

e[1996]).

Yet,m

oreis

atstake.

The

rhetoric

and

symbolism

ofaffirm

ativeaction

may

affectth

elevelof

socialexclusion

r.On

the

one

han

d,L

oury

[1995]argu

esth

atportrayin

gA

frican-A

merican

sas

vic-tim

s,aportrayaln

ecessaryto

retainaffirm

ativeaction

programs,

iscostly

toblacks.

Interm

sof

the

model,

such

rhetoric

will

increase

ran

dth

eadoption

ofR

ediden

tities.On

the

other

han

d,affirm

ativeaction

will

decreaser,

toth

eexten

tit

isseen

asan

apologyfor

previous

discrimin

ationan

dan

invitation

forblack

admission

toth

edom

inan

tcu

lture.R

eversalofaffirmative

actionw

ould

negate

this

effect.To

citea

recent

example,

our

analysis

suggests

that

removin

gaffirm

ativeaction

admission

scriteria

atth

eU

niversity

ofC

alifornia

and

University

ofTexas

Law

Sch

oolscou

ldhave

behavioral

implication

sth

atfar

exceedth

eim

pacton

applicants.

The

identity

model

ofexclu

sion,

then

,explain

sw

hy

legalequ

alitym

aynot

been

ough

toelim

inate

racialdisparities. 61

If

60.O

gbu[1997]

and

Delpit

[1995]find

that

African

-Am

ericanstu

dents

inpoor

neigh

borhoods

may

beam

bivalent

about

learnin

gStan

dardE

nglish

,w

hose

use

may

becon

strued

as‘‘actin

gw

hite.’’

61.W

esee

this

distinction

inth

edifferen

tcon

clusion

softw

orecen

tstu

diesof

U.S

.racerelation

s.Thern

stroman

dT

hern

strom[1997]u

rgean

end

toaffirm

ativeaction

,makin

gth

ecase

that

attitudes

ofwhites

toward

blacksas

wellas

the

legalopportu

nities

forblacks

have

chan

gedsin

ceT

he

Am

ericanD

ilemm

a[M

yrdal1944].

Incon

trast,Shipler

[1997]poin

tsou

tth

em

any

ways

inw

hich

African

-A

merican

san

dw

hites

feeluncom

fortablew

itheach

other

and

how

blacksare

stillseen

asdifferen

tan

dnot

fully

accepted.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

744

African

-Am

ericans

choose

tobe

Red

because

ofexclu

sionan

dif

whites

perpetuate

such

exclusion

s,evenin

legalw

ays,there

canbe

aperm

anen

tequ

ilibrium

ofracial

inequ

ality.T

he

negative

externalities

and

their

consequ

ences,

how

ever,w

ould

disappearw

hen

the

comm

unity

isfu

llyin

tegratedin

toth

edom

inan

tcu

l-tu

re,soth

atr

�a

�0,an

deveryon

ein

the

comm

unity

adoptsa

Green

identity.

This,

ofcou

rse,is

the

Am

ericanideal

ofth

em

elting

pot,orth

enew

idealof

am

osaicw

here

difference

canbe

main

tained

with

inth

edom

inan

tcu

lture.

VI.

IDE

NT

ITY

AN

DT

HE

EC

ON

OM

ICS

OF

TH

EH

OU

SE

HO

LD

An

identity

model

ofth

ehou

sehold,u

nlike

previous

models,

predictsan

asymm

etricdivision

oflabor

between

husban

dsan

dw

ives.T

heories

basedon

comparative

advantage

(e.g.,B

ecker[1965]

and

Min

cer[1962])

predictth

atw

hoever

works

more

outside

the

hom

ew

illwork

lessin

sideth

ehom

e,wheth

erit

beth

ehusban

dor

the

wife.

Yet,th

edata

we

present

belowin

dicatea

gender

asymm

etry.W

hen

aw

ifew

orksm

orehou

rsou

tsideth

ehom

e,she

stillundertakes

alarger

share

ofthe

hou

sework.

Hoch

schild’s

[1990]stu

dyT

he

Secon

dS

hift

revealsth

ede-

tailsof

such

asymm

etries.One

ofth

ecou

plesin

her

study

found

anin

geniou

sw

ayto

share

the

hou

sework.

‘‘Evan

Holt,’’a

furn

i-tu

resalesm

an,took

careof

the

lower

half

ofth

ehou

se(i.e.,

the

basemen

tan

dhis

tools).H

isw

ife‘‘N

ancy,’’a

full-tim

elicen

sedsocial

worker,

tookcare

ofth

eupper

half.

She

tookcare

ofth

ech

ild.He

tookcare

ofthe

dog.Q

uan

titativeeviden

cefrom

Hoch

schild’s

sample

and

our

dataan

alysissu

ggestth

atth

eH

oltscon

formto

anation

alpattern

.F

igure

IIIsh

ows

the

lowaverage

ofhusban

ds’share

ofhou

sework

and

itslow

elasticityw

ithrespect

toth

eirsh

areof

outside

work

hou

rs.The

figu

replots

shares

ofhou

sework

reportedby

married

men

62inth

eP

anelS

tudy

ofIncom

eD

ynam

ics, 63ascom

puted

from

62.M

en’s

reportsof

hou

sework

shares

match

edalm

ostexactly

wom

en’s

reportsin

Preston

’s[1997]stu

dyof1700

scientists.

63.T

he

unit

ofobservation

isa

couple-year

forth

eyears

1983to

1992.C

ouples

were

inclu

dedin

agiven

year,ifth

eyw

erem

arried,neith

erm

ember

was

retired,neith

erm

ember

was

disabled,the

couple

had

positivew

orkhou

rs,positiveearn

ings,an

dpositive

hou

rsofh

ousew

ork.Inaddition

,they

were

only

inclu

dedif

there

were

complete

datafrom

bothm

embers

onearn

ings,w

orkhou

rs,hou

sework

hou

rs,an

dnum

berof

children

.T

he

final

sample

had

slightly

more

than

29,000cou

ple-yearsof

observations.W

edefi

ne

ahusban

d’ssh

areof

hou

sework,h

swk,as

his

share

ofth

etotal

performed

byth

ecou

ple.T

hus,

we

capture

the

divisionof

laboreven

inhou

seholds

that

hire

outside

workers.

We

estimate

the

followin

gTobit

equation

:hsw

k�

a�

�i�

1,2,3 [b1i h

i�

b2i h

i 2�

b3i h

i 3�

b4i h

i 4]�

error,where

hi

EC

ON

OM

ICS

AN

DID

EN

TIT

Y745

answ

ersto

the

question

(s):‘‘A

bout

how

much

time

doyou

(your

wife)spen

don

hou

sework

inan

averagew

eek?I

mean

time

spent

cooking,

cleanin

g,an

ddoin

goth

erw

orkarou

nd

the

hou

se?’’The

inten

tof

the

question

was

toexclu

dech

ildcare.T

he

figu

replots

men

’ssh

areof

hou

sework

asa

fourth

-orderpolyn

omial

ofth

eirsh

areof

outside

hou

rs,for

hou

seholds

byage

ofyou

ngest

child.

When

men

doall

the

outside

work,

they

contribu

teon

averageabou

t10

percentofh

ousew

ork.Butas

their

share

ofoutside

work

falls,their

share

ofhou

sework

risesto

no

more

than

37percen

t.As

show

nin

the

figu

reth

epresen

ceof

children

ofdifferen

tages

isth

ehusban

d’ssh

areof

outside

hou

rsw

orkedif

ingrou

pi.T

he

sum

mation

(i�

1,2,3)ru

ns

overth

reetypes

ofhou

sehold:w

ithno

children

oryou

ngest

child

overage

13,with

youngest

child

0to

5,and

with

youngest

child

6to

13.Con

trolsw

erein

cluded

forage

ofhusban

d,an

dw

iferelative

topopu

lationaverage,

logof

totalin

come,

and

alsototal

hou

rsof

hou

sework.

Resu

ltsw

ererobu

stto

different

specification

san

destim

ators,an

dsu

bstitution

ofsh

areof

earnin

gsfor

share

ofhou

rsw

orked.The

equation

san

dcon

fiden

cein

tervalsare

availableupon

request.

FIG

UR

EIII

Husban

ds’Share

ofHou

sework

Hou

rsversu

sT

heir

Share

ofOutside

Work

Hou

rs

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

746

makes

asm

alldifference

toth

efu

nction

. 64Sim

ilarresu

ltsobtain

when

the

indepen

dent

variableis

shares

ofin

come

rather

than

shares

ofoutside

work

hou

rs.E

xisting

theories

donot

predictth

isasym

metry.C

onsider

the

followin

gvarian

tbased

oncom

parativeadvan

tage.Husban

dan

dw

ifeboth

have

the

same

utility

function

,w

hich

isin

creasing

inqu

antity

ofa

hou

sehold

public

goodth

atderives

fromth

eirjoin

tlabor. 65

Utility

isdecreasin

gin

own

laborin

puts

inou

tsidean

dhom

eprodu

ction. 66

We

assum

eequ

albargain

ing

power,

soth

ateach

marriage

partner

enjoys

the

same

levelofutility. 67

With

this

framew

ork,retu

rns

tospecialization

explainth

eobserved

divi-sion

oflabor

when

aw

ifehas

acom

parativeadvan

tagein

hom

eprodu

ction.W

omen

who

put

inless

than

half

ofth

eou

tsidew

orkhou

rspu

tin

more

than

half

the

hou

sework,

asseen

inth

erigh

t-han

dside

ofth

egraph

ofF

igure

III.B

ut

this

model

isin

consisten

tw

ithth

eleft-h

and

sideofth

egraph

.Iden

titycon

siderations

canexplain

the

high

shares

ofhou

se-w

orkofw

ivesw

ho

undertake

alarge

share

ofoutside

work

hou

rs.A

ddto

the

abovem

odeltwo

socialcategories,‘‘men

’’and

‘‘wom

en.’’

Prescription

sdictate

that

‘‘men

’’shou

ldnot

do‘‘w

omen

’sw

ork’’inth

ehom

ean

d‘‘m

en’’sh

ould

earnm

oreth

anth

eirw

ives.H

ochs-

child’s

interview

ssu

ggestth

atm

any

men

,and

some

wom

en,h

oldth

eseprescription

s.In

the

amen

dedm

odel,th

ehusban

dloses

identity

when

he

doeshou

sework

and

when

his

wife

earns

more

than

half

the

hou

sehold

incom

e.E

quality

ofutility

isrestored

when

the

wife

undertakes

more

hou

sework

than

her

husban

d.H

ochsch

ildreports

that

inth

e‘‘Tan

agawa’’

hou

sehold,

forex-

ample,‘‘N

ina’’earn

edm

oreth

anhalf

the

family

incom

e,but

she

64.H

erschan

dStratton

[1994]use

the

PSID

tostu

dyw

heth

erhusban

ds’high

erw

agein

comes

account

forth

eirlow

ersh

aresofh

ousew

ork.The

estimation

here,in

contrast,evalu

atesth

easym

metry

inth

erelation

ship

between

husban

ds’sh

areof

incom

ean

dth

eirsh

aresof

hou

sework,

and

wives’sh

aresof

incom

ean

dhou

sework.

65.T

he

public

goodsaspectofa

marriage

follows

Lundberg

and

Pollak

[1993],w

here

the

contribu

tions

ofeach

spouse

arein

‘‘separatesph

eres’’th

atrefl

ectgen

derroles.T

he

first

bargainin

gm

odelsofth

ehou

sehold

aredu

eto

Man

seran

dB

rown

[1980]and

McE

lroyan

dH

orney

[1981].66.

Utility

ofth

ew

ifeis

Uf

�U

f (g,hf h,h

f o),w

here

gis

the

hou

sehold

public

good,produ

cedby

bothhom

ean

dou

tsidelabor,

hf h

isth

ew

ife’shou

rsof

hou

se-w

ork,an

dh

f ois

her

outside

work.

The

husban

d’sutility

function

is,sim

ilarly,U

m�

Um

(g,hm h,h

m o),w

here

Uf an

dU

mare

assum

edto

beth

esam

efu

nction

s.67.

We

assum

eth

ata

hou

sehold

maxim

izesth

esu

mofu

tilitiessu

bjectto

the

condition

Uf�

Um

.When

bargainin

gpow

erderives

fromearn

ing

capabilitiesan

dcon

trolof

finan

cialresou

rces,as

assum

edby

Hersch

and

Stratton

[1994]an

doth

ers,itonly

reinforces

the

conclu

sionth

atwhoever

works

more

outside

the

hom

ew

orksless

inside.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y747

worked

more

than

‘‘Peter’’at

hom

eto

assuage

his

unease

with

the

situation

.Even

tually,sh

equ

ither

job.

VII.

CO

NC

LU

SIO

N

This

papercon

sidershow

identity

affectsecon

omic

outcom

es.F

ollowin

gm

ajorth

emes

inpsych

ologyan

dsociology,

identity

inou

rm

odelsis

basedon

socialdifference.A

person’s

sense

ofselfisassociated

with

differentsocialcategories

and

how

peoplein

these

categoriessh

ould

behave.

This

simple

extension

ofth

eutility

function

could

greatlyexpan

dou

runderstan

ding

ofecon

omic

outcom

es.In

aw

orldof

socialdifferen

ce,on

eof

the

most

impor-

tantecon

omic

decisions

that

anin

dividualm

akesm

aybe

the

typeof

personto

be.L

imits

onth

isch

oicew

ould

alsobe

criticaldeterm

inan

tsofecon

omic

behavior,opportu

nity,an

dw

ell-being.

Identity

affectsecon

omic

behavior

inou

rm

odelsth

rough

four

avenues.

First,

identity

chan

gesth

epayoffs

fromon

e’sow

naction

s.We

capture

this

possibilityby

avalu

eI

s inou

rm

odels.Inou

rstu

dyof

gender

inth

ew

orkplace,for

example,

aw

oman

workin

gin

a‘‘m

an’s’’job

suffersa

lossin

utility,affectin

gth

elabor

supply.S

econd,iden

titych

anges

the

payoffsofoth

ers’actions.W

ecaptu

reth

isextern

alityby

avalu

eI

o inou

rm

odels.A‘‘R

ed’’inou

rpoverty

model,

forexam

ple,is

harm

edby

am

ember

ofhis

own

comm

unity

who

complies

with

the

domin

ant

cultu

re.T

hird,

the

choice,or

lackth

ereof,ofdifferentiden

titiesaffects

anin

dividual’s

econom

icbeh

avior.Inou

rpoverty

model,w

hile

individu

alscou

ldch

oosebetw

eenG

reenor

Red,th

eycou

ldnever

bea

‘‘true’’G

reen.

The

greaterth

eexten

tof

this

socialexclu

sion,

the

greaterth

epossibility

ofequilibria

inw

hich

individu

alsesch

ewrem

unerative

activities.F

inally,

the

socialcategories

and

behavioral

prescrip-tion

scan

bech

anged,

affecting

identity-based

preferences.

This

possibilityexpan

dedth

escope

ofemploym

entpolicy

inou

rm

odelofgen

derin

the

workplace

and

ofeducation

policyin

our

study

ofsocialexclu

sion.

This

paperhas

only

scratched

the

surface

ofth

eecon

omic

implication

sof

identity.A

first

tackin

futu

reresearch

wou

ldbe

contin

ued

analysis

ofparticu

larsettin

gs.Iden

tityis

likelyto

affectecon

omic

outcom

es,for

example,

inareas

ofpolitical

econom

y,organ

izational

behavior,

demograph

y,th

eecon

omics

oflan

guage,violen

ce,education

,consu

mption

and

savings

behavior,

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

748

retiremen

tdecision

s,an

dlabor

relations. 68

As

inth

ispaper,

models

that

incorporate

well-docu

men

tedexistin

gsocial

catego-ries

and

prescriptions

could

yieldnew

results.

Asecon

dtack

inth

isagen

dais

comparative,exam

inin

giden

tityacross

spacean

dtim

e. 69R

esearchers,

forexam

ple,cou

ldcon

siderw

hy

notion

sof

‘‘class’’or

‘‘race’’vary

acrosscou

ntries;

why

migh

tgen

deran

dracialin

tegrationvary

acrossin

dustries;w

hat

migh

texplain

the

risean

dfall

ofeth

nic

tension

s.Such

comparative

studies

wou

ldbe

afru

itful

way

toexplore

the

formation

ofiden

tity-basedpreferen

ces. 70In

peroration,th

ispaper

exploreshow

toin

corporateiden

tityin

toecon

omic

models

ofbeh

avior.M

any

standard

psychological

and

sociologicalcon

cepts—self-im

age,id

ealtype,

in-grou

pan

dou

t-group,social

category,iden

tification

,anxiety,self-d

estruction

,self-realization

,situation

—fit

natu

rallyin

our

framew

ork,allow-

ing

anexpan

dedan

alysisof

econom

icou

tcomes.T

his

framew

orkis

then

perhaps

one

way

toin

corporatem

any

differentnon

pecuni-

arym

otivations

forbeh

aviorin

toecon

omic

reasonin

g,with

consid-

erablegen

eralityan

da

comm

onth

eme.

UN

IVE

RSIT

YO

FC

AL

IFO

RN

IAA

TB

ER

KE

LE

YA

ND

TH

EB

RO

OK

ING

SIN

ST

ITU

TIO

N

UN

IVE

RSIT

YO

FM

AR

YL

AN

DA

TC

OL

LE

GE

PA

RK

RE

FE

RE

NC

ES

Akerlof,G

eorgeA

.,‘‘The

Econ

omics

ofCaste

and

ofthe

Rat

Race

and

Oth

erW

oeful

Tales,’’Qu

arterlyJ

ourn

alofEcon

omics,X

C(N

ovember,1976),599–617.

,‘‘S

ocialD

istance

and

Social

Decision

s,’’E

conom

etrica,L

XV

(Septem

ber1997),1005–1027.

Alten

baugh

,R

ichard

J.,T

he

Teacher’s

Voice:A

Social

History

ofTeach

ing

inTw

entieth

-Cen

tury

Am

erica(L

ondon

:The

Falm

erP

ress,1992).A

nderson

,E

lijah,

StreetW

ise:R

ace,C

lass,an

dC

han

gein

anU

rbanC

omm

un

ity(C

hicago:U

niversity

ofChicago

Press,1990).

Andreon

i,Jam

es,‘‘G

iving

with

Impu

reA

ltruism

:A

pplications

toC

harity

and

Ricardian

Equ

ivalence,’’

Jou

rnal

ofP

oliticalE

conom

y,X

CV

II(D

ecember

1989),1447–1458.

68.See

apreviou

sversion

ofth

epaper

forsh

ortversion

sof

man

yof

these

applications.

69.W

eare

grateful

toan

anon

ymou

sreferee

forth

islist

ofcom

parativestu

dies.70.

Som

esch

olarshave

studied

the

formation

ofiden

tity-basedpreferen

cesfrom

principles

ofoptim

ization.T

heories

ofevolu

tionary

psychology,for

example,

positth

athostility

toward

‘‘outsiders’’m

ight

bein

heren

tto

hum

annatu

reas

aresu

ltof

anevolu

tionary

process;su

rvivaldepen

dson

cooperationw

ithin

sidersan

dhostility

toward

outsiders.A

noth

erfu

nction

alexplanation

ofiden

tityderives

fromsocialcogn

itionth

eory:stereotypessu

mm

arizein

formation

and

compen

satefor

hum

anbein

gs’limited

cognitive

abilities.(S

eeW

etherell

[1998,pp.

186–197]for

review.)

These

theories,

how

ever,m

ayfind

itdifficu

ltto

accomm

odateth

ecom

plexityof

socialcategories

and

prescriptions

and

the

varietyof

socialcatego-

riesacross

societiesan

dacross

time.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y749

Arrow

,K

enneth

J.,‘‘M

odelsof

JobD

iscrimin

ation’’

and

‘‘Som

eM

athem

aticalM

odelsof

Race

Discrim

ination

inth

eL

aborM

arket,’’inA

nth

ony

H.P

ascal,ed.,

Racial

Discrim

ination

inE

conom

icL

ife(L

exington

,M

A:

D.

C.

Heath

,1972).

Baldw

in,Jam

es,Th

eF

ireN

extTim

e(N

ewYork:T

he

DialP

ress,1962).B

ecker,G

aryS.,

‘‘AT

heory

ofth

eA

llocationof

Tim

e,’’Econ

omic

Jou

rnal,

LX

XV

(Septem

ber1965),493–517.

,T

he

Econ

omics

ofD

iscrimin

ation,

second

edition(C

hicago:

University

ofC

hicago

Press,1971).

Ben

abou,

Rolan

d,‘‘W

orkings

ofa

City:

Location

,E

ducation

,an

dP

roduction

,’’Q

uarterly

Jou

rnalofE

conom

ics,CV

III(A

ugu

st1993),619–652.

,‘‘H

eterogeneity,

Stratifi

cation,an

dG

rowth

:M

acroeconom

icIm

plications

ofC

omm

unity

Stru

cture

and

Sch

oolF

inan

ce,’’A

merican

Econ

omic

Review

,L

XX

XV

I(Ju

ne

1996),584–609.B

ergman

n,

Barbara

R.,

‘‘Occu

pational

Segregation

,W

agesan

dP

rofits

When

Em

ployersD

iscrimin

ateby

Race

orSex,’’E

asternE

conom

icsJ

ourn

al,I(1974),103–110.

Bhabh

a,Hom

i,‘‘Differen

ce,Discrim

ination

,and

the

Discou

rseof

Colon

ialism,’’in

Th

eP

oliticsofT

heory,F.B

arker,ed.(Lon

don:C

olchester,1983).

Blau

,Fran

cine

D.,an

dM

arianne

A.F

erber,Th

eE

conom

icsof

Wom

en,M

en,an

dW

ork(E

nglew

oodC

liffs,NJ,P

rentice-H

all,1986).B

lau,Francine

D.,P

atriciaSim

pson,andD

eborahA

nderson,‘‘Continuing

Progress?

Trends

inO

ccupation

alSegregation

inth

eU

nited

States

overth

e1970’s

and

1980’s,’’NB

ER

Workin

gP

aperN

o.6716,1998.B

owles,

Sam

uel,

and

Herbert

Gin

tis,‘‘O

ptimal

Paroch

ialism:

The

Dyn

amics

ofTru

stan

dE

xclusion

inC

omm

unities,’’m

imeo,U

niversity

ofMassach

usetts

atA

mherst,1997.

Breger,

Lou

is,F

romIn

stinct

toId

entity:

Th

eD

evelopmen

tof

Person

ality(E

ngle-

wood

Cliffs,N

J:Pren

tice-Hall,1974).

Brock,

William

A.,

and

Steven

N.D

urlau

f,‘‘D

iscreteC

hoice

with

Social

Interac-

tions

I:Theory,’’N

BE

RW

orking

Paper

No.5291,1995.

Brow

n,

Roger

W.,

Social

Psych

ology:T

he

Secon

dE

dition

(New

York:T

he

Free

Press,1986).

Bryk,A

nth

ony

S.,V

alerieE

.Lee,an

dP

eterB

.Hollan

d,Cath

olicS

chools

and

the

Com

mon

Good

(Cam

bridge,MA

:Harvard

University

Press,1993).

Bulow

,Jeremy

I.,and

Law

rence

H.S

um

mers,‘‘A

Theory

ofD

ual

Labor

Markets

with

Application

toIndustrialP

olicy,Discrim

inationand

Keynesian

Unem

ploy-m

ent,’’J

ourn

alofLabor

Econ

omics,IV

(July

1986),376–415.B

um

iller,Elisabeth

,May

YouB

eth

eM

other

ofaH

un

dred

Son

s:AJ

ourn

eyA

mon

gth

eW

omen

ofInd

ia(N

ewYork:R

andom

Hou

se,1990).B

utterfi

eld,F

ox,A

llG

od’s

Ch

ildren

:T

he

Bosket

Fam

ilyan

dth

eA

merican

Tradition

ofViolen

ce(N

ewYork:A

vonB

ooks,1995).C

lark,Ken

neth

,Dark

Gh

etto(N

ewYork:H

arper&

Row

,1965).C

ole,H

aroldL

.,G

eorgeJ.

Mailath

,an

dA

ndrew

Postlew

aite,‘‘S

ocialN

orms,

Savin

gsB

ehavior

and

Grow

th,’’J

ourn

alof

Political

Econ

omy,

C(D

ecember

1992),1092–1125.C

omer,Jam

esP.,S

choolP

ower:Im

plications

ofanIn

tervention

Project(N

ewYork:

The

Free

Press,1980).

Davies,

Margery,

Wom

en’s

Place

Isat

the

Typewriter:

Office

Work

and

Office

Workers,1870–1930

(Philadelph

ia,PA:Tem

pleU

niversity

Press,1982).

Davis,

Fred,

Th

eN

ursin

gP

rofession:

Five

Sociological

Essays

(New

York:Joh

nW

iley&

Son

s,1966).de

Grazia,

Victoria,

Th

eS

exof

Th

ings:

Gen

der

and

Con

sum

ptionin

Historical

Perspective

(Berkeley:U

niversity

ofCaliforn

iaP

ress,1996).D

elpit,L

isa,O

ther

People’s

Ch

ildren

:C

ultu

ralC

onfl

ictin

the

Classroom

(New

York:The

New

Press,1995).

Dicken

s,W

illiamT.,

and

Thom

asJ.

Kan

e,‘‘R

acialan

dE

thnic

Preferen

cein

College

Adm

issions,’’

Brookin

gsP

olicyB

riefN

o.9

(Wash

ington

,D

C:

The

Brookin

gsIn

stitution

,Novem

ber1996).

DuB

ois,W

illiamE

.,T

he

Sou

lsof

Black

Folk

(Green

wich

,C

T:

Faw

cettP

ublica-

tions,1965).

Elster,

Jon,

Th

eC

emen

tof

Society:

AS

tud

yof

Social

Ord

er(C

ambridge:

Cam

-bridge

University

Press,1989).

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

750

Fan

on,F

rantz,B

lackS

kin,W

hite

Masks

(New

York:Grove

Press,1967).

Fish

er,Sue,N

ursin

gW

oun

ds:N

urse

Practition

ers/D

octors/W

omen

Patien

ts/A

nd

the

Negotiation

ofM

eanin

g(N

ewB

runsw

ick,N

J:R

utgers

University

Press,

1995).F

liegel,Seym

our,

Miracle

inE

astH

arlem:

Th

eF

ight

forC

hoice

inP

ublic

Ed

ucation

(New

York:Ran

domH

ouse,1993).

Folbre,

Nan

cy,W

ho

Pays

forth

eK

ids:

Gen

der

and

the

Stru

ctures

ofC

onstrain

t(N

ewYork:R

outledge,1994).

Fran

ke,K

atherin

eM

.,‘‘T

he

Cen

tralM

istakeof

Sex

Discrim

ination

Law

:T

he

Disaggregation

ofSex

fromG

ender,’’U

niversity

ofP

enn

sylvania

Law

Review

,C

XL

IV(N

ovember

1995),1–99.F

razier,Fran

klin,T

he

Black

Bou

rgeoisie:Th

eR

iseofth

eN

ewM

idd

leC

lassin

the

Un

itedS

tates(N

ewYork:T

he

Free

Press,1957).

Friedm

an,M

ilton,

Essays

inP

ositiveE

conom

ics(C

hicago:

University

ofC

hicago

Press,1953).

Fulw

oodIII,S

am,W

aking

fromth

eD

ream:M

yL

ifein

the

Black

Mid

dle

Class

(New

York:Dou

bleday,1996).G

andh

i,Moh

andas,A

utobiograph

y(L

ondon

:Jonath

onC

ape,1966).G

erson,

Kath

leen,

Hard

Ch

oices:H

owW

omen

Decid

eabou

tW

ork,C

areer,an

dM

otherh

ood(B

erkeley,CA

:University

ofCaliforn

iaP

ress,1986).G

leitman

,Hen

ry,Basic

Psych

ology(N

ewYork:N

orton,1996).

Goldin

,Clau

dia,Un

derstan

din

gth

eG

end

erG

ap:An

Econ

omic

History

ofA

meri-

canW

omen

(New

York:Oxford

University

Press,1990a).

,‘‘AP

ollution

Theory

ofDiscrim

ination

,’’mim

eo,Harvard

University,1990b.

Goodson

,IvorF.,an

dA

ndy

Hargreaves,Teach

ers’Profession

alLives

(Lon

don:T

he

Falm

erP

ress,1996).H

annerz,U

lf,Sou

lside:In

quiries

into

Gh

ettoC

ultu

rean

dC

omm

un

ity(N

ewYork:

Colu

mbia

University

Press,1969).

Hersch

,Jon

i,an

dL

eslieS.

Stratton

,‘‘H

ousew

ork,W

ages,an

dth

eD

ivisionof

Hou

sework

Tim

efor

Em

ployedSpou

ses,’’A

merican

Econ

omic

Association

Papers

and

Proceed

ings,L

XX

XIV

(May

1994),120–125.H

ochsch

ild,Arlie,w

ithA

nne

Mach

ung,T

he

Secon

dS

hift(N

ewYork:A

von,1990).

Hon

ey,M

aureen

,C

reating

Rosie

the

Riveter:

Class,

Gen

der,

and

Propagan

da

du

ring

World

War

II(A

mherst:U

niversity

ofMassach

usetts

Press,1984).

Huan

g,Peter

H.,an

dH

oM

ouW

u,‘‘M

oreO

rderw

ithou

tL

aw:A

Theory

ofSocial

Norm

san

dO

rganization

alCultu

res,’’Jou

rnal

ofL

aw,E

conom

ics,&O

rgani-

zation,X

(October

1994),390–406.K

andori,

Mich

ihiro,

‘‘Social

Norm

san

dC

omm

unity

Enforcem

ent,’’

Review

ofE

conom

icS

tud

ies,LX

IX(Jan

uary

1992),63–80.K

anter,

Rosabeth

Moss,

Men

and

Wom

enof

the

Corporation

(New

York:B

asicB

ooks,1977).K

evane,

Mich

ael,‘‘C

anT

here

Be

an‘Iden

tityE

conom

ics’?’’m

imeo,

Harvard

Academ

yfor

Intern

ationalan

dA

reaStu

dies,1994.K

hatibi,A

bdelkebir,La

Blessu

red

uN

omP

ropre(P

aris:Den

oel,1986).K

hattab,

Hin

d,‘‘W

omen

’sP

erceptions

ofSexu

alityin

Rural

Giza,’’R

eproductive

Health

Workin

gG

roup,T

he

Popu

lationC

ouncil,M

onograph

sin

Reprodu

ctiveH

ealthN

o.1,1996.K

uran

,Tim

ur,

‘‘Eth

nic

Norm

san

dT

heir

Transform

ationth

rough

Repu

tationC

ascades,’’Jou

rnalofL

egalStu

dies,X

XV

II(S

um

mer

1998,Part

2),623–659.L

anda,

Janet

T.,Tru

st,E

thn

icity,an

dId

entity:

Beyon

dth

eN

ewIn

stitution

alE

conom

icsof

Tradin

gN

etworks

(Ann

Arbor:

University

ofM

ichigan

Press,

1994).L

azear,E

dward

P.,an

dSherw

inR

osen,‘‘M

ale-Fem

aleW

ageD

ifferentials

inJob

Ladders,’’J

ourn

alofLabor

Econ

omics,V

III(Jan

uary

1990),S106–S

123.L

evine,

Judith

A.,

‘‘It’sa

Man

’sJob,

orSo

They

Say:

The

Produ

ctionof

Sex

Segregation

inO

ccupation

s,’’mim

eo,Departm

ent

ofSociology,N

orthw

esternU

niversity,1995.

Liebow

,E

lliott,T

ally’sC

orner:

AS

tud

yof

Negro

Streetcorn

erM

en(B

oston:

Little-B

rown,1967).

Loew

enstein

,George,‘‘B

ecause

ItIs

There:T

he

Challen

geof

Mou

ntain

eering

...for

Utility

Theory,’’m

imeo,C

arnegie

Mellon

University,1998.

Lou

ry,Glen

nC

.,On

eby

On

efrom

the

Insid

eO

ut(N

ewYork:T

he

Free

Press,1995).

EC

ON

OM

ICS

AN

DID

EN

TIT

Y751

Lundberg,

Shelly,

and

Robert

A.

Pollak,

‘‘Separate

Sph

eresB

argainin

gan

dth

eM

arriageM

arket,’’Jou

rnal

ofP

oliticalE

conom

y,C

I(D

ecember

1993),988–

1010.M

acKin

non

,C

atharin

eA

.,S

exual

Harassm

ent

ofW

orking

Wom

en(N

ewH

aven,

CT:Yale

University

Press,1979).

MacL

eod,Jay,Ain

’tN

oM

akin’It:L

eveledA

spirations

ina

Low

-Incom

eN

eighbor-

hood

(Bou

lder,CO

:Westview

Press,1987).

Man

ser,Marilyn

,and

Murray

Brow

n,‘‘M

arriagean

dH

ouseh

oldD

ecisionM

aking:

AB

argainin

gA

nalysis,’’In

ternation

alEcon

omic

Review

,XX

I(F

ebruary

1980),31–44.

Mason

,K

arenO

ppenheim

,‘‘C

omm

entary:

Strober’s

Theory

ofO

ccupation

alSex

Segregation

,’’C

hapter

9in

Barbara

Reskin

,ed.,

Sex

Segregation

inth

eW

orkplace(W

ashin

gton,D

C:N

ationalA

cademy

Press,1984).

McE

lroy,M

arjorieB

.,an

dM

aryJean

Horn

ey,‘‘N

ashB

argained

Hou

sehold

Decision

s,’’Intern

ationalE

conom

icR

eview,X

XII

(June

1981),559–583.M

eier,D

eborah,

Th

eP

ower

ofT

heir

Ideas:

Lesson

sfor

Am

ericafrom

aS

mall

Sch

oolinH

arlem(B

oston,M

A:B

eaconP

ress,1995).M

ilkman

,Ruth

,Gen

der

atW

ork:Th

eD

ynam

icsof

Job

Segregation

byS

exd

urin

gW

orldW

arII

(Urban

a:University

ofIllinois

Press,1987).

Miller,

Kerby

A.,

‘‘Assim

ilationan

dA

lienation

:Irish

Em

igrants’

Respon

sesto

Indu

strialA

merica,’’in

P.J.

Dru

dy,ed.,

Th

eIrish

inA

merica:

Em

igration,

Assim

ilationan

dIm

pact(C

ambridge,

UK

:C

ambridge

University

Press,

1985).M

incer,

Jacob,an

dSolom

onP

olachek,

‘‘Fam

ilyIn

vestmen

tsin

Hum

anC

apital:E

arnin

gsof

Wom

en,’’J

ourn

alof

Political

Econ

omy,

LX

XX

II(M

arch1974),

S76–S

108.M

incer,

Jacob,‘‘L

aborF

orceP

articipationof

Married

Wom

en:A

Stu

dyof

Labor

Supply,’’in

Aspects

ofLabor

Econ

omics,C

onferen

ceN

o.14ofth

eU

niversities-

Nation

alBureau

Com

mittee

forE

conom

icR

esearch(P

rinceton

,NJ:P

rinceton

University

Press,1962).

Mon

tgomery,Jam

esD

.,‘‘Revisitin

gT

ally’sC

orner:M

ainstream

Norm

s,Cogn

itiveD

issonan

cean

dU

nderclass

Beh

avior,’’R

ationality

and

Society,

VI

(1994),462–488.,

‘‘Towards

aR

ole-Theoretic

Con

ceptionof

Em

beddedness,’’m

imeo,

Lon

donSch

oolofEcon

omics,1997.

Mottu

s,Jane

E.,N

ewYork

Nigh

tingales:T

he

Em

ergence

ofthe

Nu

rsing

Profession

atB

ellevue

and

New

YorkH

ospital1850–1920

(Ann

Arbor:

University

ofM

ichigan

Press,1981).

Myrdal,

Gunnar,

An

Am

ericanD

ilemm

a:T

he

Negro

Problem

and

Am

ericanD

emocracy

(New

York:Harper

and

Row

,1944).N

elson,

Jill,Volu

nteer

Slavery:

An

Au

then

ticN

egroE

xperience

(New

York:P

engu

in,1993).

Nisbett,R

ichard

E.,an

dD

ovC

ohen

,Cu

lture

ofHon

or:Th

eP

sychology

ofViolen

cein

the

Sou

th(B

oulder,C

O:W

estviewP

ress,1996).N

orton,A

nne,R

eflection

son

PoliticalId

entity

(Baltim

ore,MD

:The

Johns

Hopkin

sU

niversity

Press,1988).

Ogbu

,Joh

nU

.,‘‘B

eyond

Lan

guage:

Ebon

ics,P

roperE

nglish

and

Identity

ina

Black

Am

ericanSpeech

Com

munity,’’

mim

eo,U

niversity

ofC

alifornia

atB

erkeley,Departm

entofA

nth

ropology,1997.O

kuno-F

ujiw

ara,M

.,an

dA

ndrew

Postlew

aite,‘‘S

ocialN

orms

and

Ran

domM

atchin

gG

ames,’’G

ames

and

Econ

omic

Beh

avior,IX(A

pril1995),79–109.P

adavic,Iren

e,‘‘T

he

Re-C

reationof

Gen

derin

aM

aleW

orkplace,’’S

ymbolic

Interaction

,XIV

(1991),279–294.P

ierce,Jen

nifer,

Gen

der

Trials:E

motion

alL

ivesin

Con

temporary

Law

Firm

s(B

erkeley:University

ofCaliforn

iaP

ress,1995).P

ierson,R

uth

Roach

,T

hey’re

Still

Wom

enA

fterA

ll:T

he

Secon

dW

orldW

aran

dC

anad

ianW

oman

hood

(Toronto:M

cClellan

dan

dStew

art,1989).P

reston,

Anne,

‘‘Sex,

Kids,

and

Com

mitm

ent

toth

eW

orkplace:E

mployers,

Em

ployees,and

the

Mom

my

Track,’’RussellS

ageF

oundation

Workin

gP

aperN

o.123,1997.P

ringle,R

osemary,S

ecretariesT

alk:Sexu

ality,Pow

eran

dW

ork(N

ewYork:V

erso,1988).

Rabin

,M

atthew

,‘‘In

corporating

Fairn

essin

toG

ame

Theory

and

Econ

omics,’’

Am

ericanE

conom

icR

eview,L

XX

XIII

(Decem

ber1993),1281–1302.

QU

AR

TE

RL

YJ

OU

RN

AL

OF

EC

ON

OM

ICS

752

Rain

water,L

ee,Beh

ind

Gh

ettoW

alls:Black

Fam

iliesin

aF

ederalS

lum

(Chicago,

IL:A

ldine,1970).

Reskin

,Barbara,an

dP

atriciaR

oos,eds.,Job

Qu

eues,G

end

erQ

ueu

es:Explain

ing

Wom

en’s

Inroad

sIn

toM

aleO

ccupation

s(P

hiladelph

ia,PA:Tem

pleU

niversity

Press,1990).

Rodrigu

ez,Rich

ard,Hu

nger

ofMem

ory:Th

eE

du

cationofR

ichard

Rod

riguez

(New

York:Ban

tam,1982).

Roediger,D

avidR

.,Th

eW

agesofW

hiten

ess(N

ewYork:V

ersoP

ress,1991).R

omer,

Pau

lM

.,‘‘P

references,

Prom

ises,an

dth

eP

oliticsof

Enligh

tenm

ent,’’

mim

eo,University

ofCaliforn

iaat

Berkeley,D

ecember,1994.

Sach

s,Susan

,‘‘Muslim

Sch

oolsin

U.S

.aVoice

forIden

tity,’’Th

eN

ewYork

Tim

es,N

ovember

10,1998,A1.

Said,E

dward

W.,O

rientalism

(New

York:Ran

domH

ouse,1978).

San

day,Peggy

Reeves,F

raternity

Gan

gR

ape:Sex,B

rotherh

ood,an

dP

rivilegeon

Cam

pus

(New

York:New

YorkU

niversity

Press,1990).

Sch

ultz,

Vicki,

‘‘Recon

ceptualizin

gSexu

alH

arassmen

t,’’YaleL

awJ

ourn

al,C

VII

(April1998),1683–1805.

Sen

,Am

artyaK

.,‘‘T

he

Impossibility

ofa

Paretian

Liberal,’’J

ourn

alof

Political

Econ

omy,L

XX

VIII

(January

1970),152–157.,

‘‘Goals,

Com

mitm

ent,

and

Identity,’’

Jou

rnal

ofL

aw,

Econ

omics,

and

Organ

ization,I

(Fall1985),341–355.

,‘‘M

aximization

and

the

Act

ofC

hoice,’’

Econ

ometrica,

LX

V(Ju

ly1997),

745–779.Shipler,D

avid,AC

oun

tryof

Stran

gers:Blacks

and

Wh

itesin

Am

erica(N

ewYork:

Knopf,1997).

Stan

ley,M

arcus,

Law

rence

Katz,

and

Alan

Kru

eger,‘‘Im

pactsof

Em

ploymen

tP

rograms:T

he

Am

ericanE

xperience,’’m

imeo,H

arvardU

niversity,1998.

Staples,

Bren

t,P

arallelT

ime:

Grow

ing

Up

inB

lackan

dW

hite

(New

York:P

anth

eon,1994).

Stin

nett,T.M

.,Profession

alProblem

sofTeach

ers(N

ewYork:M

acmillan

,1968).Strober,

Myra,

‘‘Toward

aG

eneral

Theory

ofO

ccupation

alSex

Segregation

:T

he

Case

ofP

ublic

Sch

oolTeach

ing,’’

Chapter

8in

Barbara

Reskin

,ed.,

Sex

Segregation

inth

eW

orkplace(W

ashin

gton:N

ationalA

cademy

Press,1984).

Sugg,

Reddin

gS.,

Moth

erteacher:

Th

eF

emin

izationof

Am

ericanE

du

cation(C

harlottesville:U

niversity

ofVirgin

iaP

ress,1978).Suskin

d,Ron

,AH

opein

the

Un

seen(N

ewYork:B

roadway,1998).

Tajfel,Hen

ri,and

John

Turn

er,‘‘An

Integrative

Theory

ofIn

tergroup

Con

flict,’’in

Th

eS

ocialPsych

ologyofIn

tergroup

Relation

s,William

G.A

ustin

and

Steph

enW

orchel,eds.(M

onterey,C

A:W

adsworth

,1979).T

hern

strom,S

tephan

,and

AbigailT

hern

strom,A

merica

inB

lackan

dW

hite:O

ne

Nation

,Ind

ivisible(N

ewYork:S

imon

and

Sch

uster,1997).

Thom

as,Kerry,‘‘T

heD

efensiveSelf:A

Psychodynam

icP

erspective,’’inU

nd

erstand

-in

gth

eS

elf,R

ichard

Steven

s,ed.

(Thou

sand

Oaks,

CA

:SA

GE

Publication

s,1996).

Turow

,Scott,O

ne

L(N

ewYork:W

arner

Books,1977).

Watkin

s,Mel,D

ancin

gW

ithS

trangers:A

Mem

oir(N

ewYork:S

imon

and

Sch

uster,

1998).W

etherell,

Margaret,

‘‘Grou

pC

onflict

and

the

Social

Psych

ologyof

Racism

,’’inG

roup

Con

flict

and

the

Psych

ologyof

Racism

,M

argaretW

etherell,

ed.(T

hou

sand

Oaks,C

A:S

AG

EP

ublication

s,1996).W

hyte,

William

Foote,

Street

Corn

erS

ociety:T

he

Social

Stru

cture

ofan

ItalianS

lum

(Chicago:U

niversity

ofChicago

Press,1943).

William

s,C

hristin

e,G

end

erD

ifferences

atW

ork:W

omen

and

Men

inN

ontrad

i-tion

alOccu

pations

(Berkeley:U

niversity

ofCaliforn

iaP

ress,1989).W

illis,Pau

lR

.,Learn

ing

toL

abour:

How

Workin

gC

lassK

ids

Get

Workin

gC

lassJ

obs(W

estmead,F

arnborou

gh,H

ants.,U

K:S

axonH

ouse,1977).

Wilson

,William

J.,Th

eTru

lyD

isadvan

taged(C

hicago,IL

:University

ofC

hicago

Press,1987).

,Wh

enW

orkD

isappears:Th

eW

orldofth

eN

ewU

rbanP

oor(N

ewYork:K

nopf,

1996).W

urzbu

rg,Lyn

ne

A.,

and

Robert

H.K

lonoff,

inH

opeL

andrin

ean

dE

lizabethA

.K

lonoff,eds.D

iscrimin

ationA

gainst

Wom

en:P

revalence,C

onsequ

ences,R

em-

edies,(T

hou

sand

Oaks,C

A:S

AG

EP

ublication

s,1997),pp.175–195.

EC

ON

OM

ICS

AN

DID

EN

TIT

Y753