American Apartheid: Segregation and the Making of the Underclass

American Apartheid Segregation and the Making of the Underclass

DOUGLAS s. MASSEY NANCY A. DENTON

Harvard University Press Cambridge, Massachusetts London, England

Copyright CJ 1993 by the President aml Fellows of Hal1lard College

All rlghts reserved

Printed In the UnJted States of America

Ninth printing, 2001

LJbrary of COngress CIIllIlDging-in-P",blicllIion DQ/Q MasKy. Douglas S.

American apartheid: segregal!on and the making of the underclass

I Douglas S. Massey. Nanty A. Denton. p. em . Includes Index.

ISBN 0-674-01820-6 (al;id·frcel (clolh) ISBN 0-674-01821-4 (pbk.1 1. Race dlscrlmlnatlon-Unlted States-Hisrory-20Ih century. 2. SegregaJlon-UnJted Slates-History-2Oth century. Americans-Social conditions. Hlslory-20th century.

5. United States-Rate relations.

6. UnlJed Slates-Social polity. I. Denton. Nanty A. E185.6I.M373 )05.8'00973-dc20

3. Afro­

4. hUler cities-United Slates­

D. Title.

1993

92-13889 elP

.

To

Dr. Kenneth B. Clark

an inspiration to a generation

Contents

Preface

viii

1

1

The Missing Link.

2

The Construction of the Ghetto

3

The Persistence of the Ghetto

4

The Continuing Causes of Segregation

5

The Creation of Underclass Communities

6

The Perpetuation of the Underclass

7

The Failure of Public Policy

8

The Future of the Ghetto

Notes

239

Index

285

17 60

186 217

83

148

115

Preface

This book grew out of rese arch [hat was originally funded by the National Institute of Child Health and Human Developm e n t in June 1984 (grant number ROI-HD- 1 8594) to undertake a systematic study of racial and ethnic segregation based on the 1 980 Census . A follow-up project to examine the consequences of residential se gre gati on was fu nded by the same ag enc y in June 1987 ( g ra nt number ROI-HD-22992). As work progressed on thes e project s, the unique s egreg a tion of black Americans stood out in ever-sharper relief, and the deleterious consequences they suffered as a re sult of this spatial is o la t ion be ca me pa infu lly obvio us . Equally obvious w as that ille se facts were ignored in ongoing debates about the relative importance of race in American society and the origins of the urban underclass. It seemed to us amazing that pe ople were even debating whether race was d eclining in importance w h en levels of res i­ dential segregation were so high and so structured along racial lines , and we di d not u nderstand how the volumes of material written on the

underclass could gloss over the p ersi sting reality of raci al segregation as jf it were irrelevant to the creation and maintenance of urban poverty. OUf research indicates that racial residential se g regat io n is the principal structural feature of American society responsible fo r the perpetuation of urban po verty and repre sen t s a primary cause of racia l inequality in the

United States.

The book and its underlying research are re ally a collaborative effort of many people and institutions. First and foremost. we owe a d ebt of thanks to the National Institute of Child He alth and Human D evel opment (and hence to U.S. taxpayers) for sponsoring most of t he research under­ lying Ihis book, and we are particularly indebted [0 Dr. V. Jeffery Evans of that instit ute for his constant help and encouragemenl. The Nat ion al Opinion Research Center administered the research project and we ac-

Preface

ix

knowledge th e dedicated efforts of its outstanding swff. We also thank [he G ugg e n he i m Foundallon for a fellowship awarded to Douglas Massey during 1990-91 t hat granted him a sabbatical year in which to prepare the manuscript, and we thank the University of Chicago and its Division of the Social Scien ces for provid ing additional salary and research SUPP0rL d uring this leave. Thanks are also owed to the New York State United University Professors for a New Faculty Development Award to Nancy Denton in 1 990-91.

a hosl of students and col­ expertise and k nowledge to research du r­ ing va rious phases of the project. Isabel Garcia and Adelle Hinojosa were responsible for day-to-d ay administration and project support. Stuart Bo­ gom, Mark Keintz, Michael Strong, and Stephen Taber pr ov ided invalu­ The project also benefited greatly from

leagu es who contributed their

able assistance in creating the computerized data files u sed in later inves­ tigat ions, and we also benefited from the hard

work of B rendan Mullan

and Fe lipe Garda. Gretchen Condran compiled a special data set co v­

ering Philadelphia and

collaborated in a stu dy of segregation's social,

economic, and health consequences in that city, and Eric Fang helped extend this

a nalysis to San F r anc isco Mitchell Eggers worked intensively .

us to understand the determinants of u rba n poverty and its spatial concentration. Andrew G ro ss carried am research on methodological is­ sues and contributed ta studies of the cau ses and consequences of black segregation. Adam Bickford collaborated in a study of racial s e gregati on within U.S. p ubl ic housing and Shawn Kanaiaupuni carried out addi­ with

,

tional research to determine the effect of project location on poverty concentratian. Richard Sander

worked with us on a theoretical and em­

of neighborhood racial transition. Katharine Donato, George Galster, Hector Cordero Guzman, Harvey Molotch, and several anon y m ou s reviewers gave gener­ ausly of their time in r ead ing all or some of the manuscript, and Elizabeth Gretz improved its prose through her careful editing. The comments and suggestions of these in dividu als greatly improved the manuscript, and they can not be faulted for the shortcomings that no doubt remain. The book also reflects the patience and support of many friends and family members, particularly Susan Ross and John Pipk in pirica l analysis

Andrew Cherlin,

.

We are

grateful to several authors and institutions for permissi on to published materials. We thank Stanley L ieberson and the University of California Press for pe r mission La use ind ices originally re por ted in Piece of the Pie: Blacks and White Immigrants since 1880, © 1980 cile or quote thei r

Preface

x

by the Regents of the University of California. We also thank Reynolds Farley, Suzanne Bianchi, Diane Colasanto, and the American Academy of Political and Social Science for permission to reprint data from "Barri­ ers to the Racial Integration of Neighborhoods: The Detroit Case," from volume 441 of The Annals of the American Academy of Political and Social Science,

© 1979 by the American Academy of Political and Social Science.

We again thank Reynolds Farley, Suzanne Bianchi, and Diane Colasanto as well as Howard Schuman, Shirley Hatchett, and Academic Press for permission to reproduce data from" 'Chocolate City, Vanilla Suburbs':

Will the Trend toward Racially Separate Communities Continue?" pub­

lished in volume 7 of Social Science Research, © 1978 by Academic Press. Finally, we express our gratitude to Roderick J. Harrison and Daniel J. Weinberg of the U.S. Bureau of the Census for sending us preliminary

calculations of 1990 segregation indices prior to their presentation in "Racial and Ethnic Residential Segregation in 1990" at the April 13, 1992, meetings of the Population Association of America in Denver, Col-

0rado. To all these people and institutions we extend our deep and heartfelt

thanks, and hope that the end result justifies the sacrifices that they made on our behalf.

Chicago, Illinois March 1992

Racial segregation, like all other forms of cruelty and tyranny, debases all human beings-those who are its victims, those who victimize, and in quite subtle ways those who are mere accessories.

Kenneth

B.

Clark

1

The Missing Linl

1 2.8

2.3

1 .2

1 3.2

1 1 .7

10.1

12.4

- 1.5

4. 7

8.8

' 4.6

I Ll

9.6 1 J .7

1 2. 6

12.1

3. 1

2.3

0.6 2.7

- 0.6

Los Angeles-Long Beach

26.7

29.3

2.6

12 . 1

1 3.7

1.6

Milwaukee

30.0

33.2

3.2

1 0.4

1 1 .7

New York

25.9

1 1 .7

1 3.8

17.7

1.3

3.9

Newark

23.4

10.8

10. 1

2.2

37. 6 34. 1

Philadelphia

26.0

8.7

10.8

St. Louis

34.7

1 2.3 13.8

30.7

33.7

2.4

12.2

1 1 .9

3.0 - 0. 3

Average

2 7.0

33. 2

6.2

1 1 .2

12.8

1 .6

Source: Douglas S. Massey and Mitchell 1. Eggers, "The Ecology of Inequality:

Minorities and the Concentration of Poverty, 1 970-1980," American Journal of Socj% gy 9 5 ( 1 990) : 1 1 7 5-77.

130

AME R I C AN APARTHEID

change-essentially the conditions w e assumed in our simulati on ex . . ercise. ' As the indices of Table 5 . 2 show, blacks experienced systematically greater concentrations of poverty than whites. In 1 9 70, for examp le, Ihe average level of black poverty concentration, 2 7 % , was nearly three times

the white level of 1 1 % : and over the ensuing decade this disparity grew,

with the average black concentration index rising to 3 3 %, an increase of 6 percentage points, while the index for whites increased only mOdestly to 1 3 % , a change of only 2 points. The increase in black poverty concen.

tration was especially ma rked in Chicago, New York, N ewark, Buffalo,

and philadelphia. By the end of the decade, the typical poor black family

in the se hypersegregated metropolitan areas lived in a neighborhood

where at least a third of the families were poor In contrast, poor white families in these areas generally lived in neighborhoods where the pov· ..

erty rate was 1 5 % or less .

In statistical terms, levels and trends in poverty concentration during the 1 970s correlate strongly with the product of the index of black-white

segregation and the rate of black poverty, suggesting a powerful interac­

tion between the two factors.24 An interaction means that the effect of

rising poverty on poverty concentration depends cruciaUy on the level of

racial segregation at which the increase in poverty occurs. B lack poverty

concentration increased most dramatically in metropolitan areas where

a sharp increase in black poverty was paired with a high and unchanging level of radal segregation. By 1 980 poverty concentration was greatest

in urban areas where blacks were both highly segregated and very poor, and was lowest in areas where they were neither poor nor segregated.

Segregation was thus the key conditioning factor responsible for the

transformation of urban b l a ck communities observed by William Julius

Wilson during the 1 9 70s. In the absence of racial segregation, the same

downward economic trends would not have produced such dramatic

changes in the concentration of poverty.

The Creation of Underdass Communities Poverty, of course, is not a neutral social factor. Associated with it are a variety of other social ills such as family instability, welfare dependency,

crime, housing abandonment. and low educational achievement. To the extent that these factors are associated with poverty, any s tructural pro­

cess that concentrates poverty will concentrate them as well. Segregation,

.

'

131

The Creation of Underdass Communities

interacting with high or rising poverty rates, guarantees that blacks will face a harsh and uniquely disadvantaged social environment, no matter

wh at their personal characteristics. s eg regation, moreover, is c ru ci a l

to

1 urban black \

understanding why a self-

perpetuating spiral of neighborhood decline is built into cOlrununities. The s ocioeconomic health of black n eigh b o rh o od s is fragile

and easi ly j olted into a p attern of decay because of a subtle interaction between individual and collective behavior that comes into play when­

eve r independent actors make decisions that affe ct the welfare of the community. This interaction occurs b ec ause a neighborhood's socioeco­

i

nomic status is. to a la rge extent, determined by the decisions and behav­

iors of its inhabitants. eco nom ic cont ext wi

ach individual dec ision affects the soc ia l and

Y

in which subsequent decisions are made, creating

a powerful feedback loop between individual and c o llec tive b eh avior . Z

Consider, for example, the case of a working-class n ei ghbo rhood that

has just made [he transition from white to bl a ck , The homes and a p a rt ­

ments are well maintained and there is no abandoned housing or physi­

cal dilapidation. But suppose the new black residents have somewhat

lower incomes tha n the former white occupants. B l ack renters can afford

less rent and some black owners are less able to make investments in

upkeep . As a res u l t , a few owner s and landlords decide to spend less money on the maintenance of their propert ies . Soon their buildings begin to show signs of disinvestment in the form of p hysical disrepair: peeling pai n t, rotting gutters, unkept yards, crumbling porches, and so on. The presence of dilapidated dwellings changes the context within which other landlords and homeowners make their decisions on whether

or not t o invest. Studies suggest that property owners are extremely sensi­

live to small signs of physic al deterioration.26 The pres ence of even a

s m al l number of dilapidated buildings is ta ken as a signal that the neigh­

borhood is going "downhill. " To the extent that property owners per­

ceive a decline as pos s ibl e or likely, they have l ittle incentive to invest in

upkeep and improvement on their own buildings, because money put

into neighborhoods that are declining is unlikely to be recouped in the

form of higher rents or greater home equity. As

a result of the i ni tial disinvestment by a few ow ner s, therefore,

others are led to cut back on the money they invest. With every addi­

tional property ow n er who decides not to invest it becomes increasingly like ly that others will reach similar decisions, even if they are Otherwise disposed to maintain their buildings. At some point, a threshold is

1 32

AM E R I C A N A P A R T H E I D

crossed, beyond which the pattern becomes self-reinforcing an d ine • .

versible.27

Racial segregation makes nei ghbo rhoo d s where blacks live particularly :' vulnerable to this sort of disinvestment and decay. Poor blacks are more likely than th e poor of any other group to be trapped in neighborho ods : ca ught in the grip of such downward spirals, because segregation acts t concentrate poverty and all things associated with it. Increa s ing the pov­

�

erty rate of a segregated group concentrates not only poverty but hOUSing:

dilapidation and abandonment.

,��ssence, segregatio� and rising poverty interact to deliver �� !0.i:ls' shock to black neIghborhoods that pushes them beyond the-pi'5ii)t __

:

where physical decay and disinvestment become self-perpetuating . The

potential strength of this shock has been demonstrated in a series of simulation exercises that are sununarized in Table 5 . 3 . Again using the

hypothetical class-segregated ci ties we described earlier. we ask what wou ld happen to black neighborhoods if black poverty rates rose under different conditions of segregation.28 At different segregation levels. of course, the shift in black poverty rates yields different concentrations of

poverty, and using empirical correlations between neighborhood p o verty levels and various indicators of conununity well-being, we predict the neighborhood environment created by a particular combination of segre­ gation and poverty. In carrying out the simulations, we consider the general case of a class-segregated city and document the effect of a shift in black poverty rates from 20% to 30%. Consider the degree of housing abandonment likely to result from an

increase in black poverty under different conditions of segregation. In

the absence of racial segregation, the poverty rate in neighbo rho ods

where poor blacks live increases from 2 5 % to 30% as a result of the

assumed shift in black poverty rat es from 20% to 30% (as was shown in

Table 5 . 1 ) . Given the empirical re la tionship between a tract's poverty

,

rate and its percentage of boarded-up dwe l l i ngs this shift in neighbor­

hood poverty is predicted to increase the incidence of boarded-up struc­ tures from 1 .2 % before the shift to 1 .5 % afterward.

Although these figures are quite small. a review carried out by the u.s.

Department of Housing and Urban Development shows that the thresh­ old of neighborhood stability occurs at remarkably low percentages of

boarded-up buildings.29 Once a few structures become abandoned, the

process of decay quickly becomes cumulative and " spreads fr om older, dilapidated sections to areas with sound housing that would be well

worth the cost of maintenance and renovation if neighborhood (ondi-

" The Creation ot Underclass Communities Table S.l

133

Simulated effects of a shift in the black poveny rate on

neighborhood conditions experienced by poor blacks, assuming different levels of radal segregation

------Effect of shift in black poveny rale

Neig hborhood condition and level of racial segregation

Poveny

Poyeny

of 20%

of 30%

Change

ra le

rate

Percentage of houses boarded up No racial segregation

l .2

1.5

0.3

2.3

3.3

1 .0

$ 1 3 ,020

$ 1 1 .235

$8, 1 60

$4.523

$ \ ,785 $3.637

No racial segregation

57.8

6 1 .8

4.0

Complete radal segregation

68. 3

84.2

1 5.9

No racial segregalion

21.1

24.6

3.5

Complete racial segregation

36. 1

5 1 .0

14.9

1 9. 2

22.2

3 .0 1 2.0

complete racial segregation

Median household income

No racial segregation

Complele racial segregation

Major crime rate (per

1, 000 people)

Percentage offamilies on welfare

Percmtage offemale- headed families

No racial segregation Complete racial segregation

33.5

45.5

Percentage of high school students scoring below 15th percentile No radal segregation

Complete racial segregation

3i.6

47. 1

35.3

2.7

57.8

1 0. 7

Source: Douglas S . Massey, "American Apanheid: Segregation a nd the Milking o f the Underciass, " Ameruan }oUTnilI of Sodolo9Y 96 ( 1990):343. 348-49.

Note: Simulalion resulls assume residential structures depicted in Figure 5 . 1 with class

segregation between poor and nonpoor families added; test rderred to In last panel is

the California Achievement Test.

tions were more stable," 3o The study concludes that a "tipping point" is reached once 3%-6% of a neighborhood's dwellings are abandoned,

"the point at which investment psycho1ogy becomes so depressed that

reversal ofthe abandonment process is impossible without major external intervention."3!

1 34

AM E R I C A N A P A RT H E I D

I n the absence of raci a l segregation, therefore, an exogenous eco nomic sh ock that causes black poverty rates to ri se does not concentrate ho uSin ­

g

ab andonment enough to reach th i s tipping point. Under cond iti ons of complete segregation, h o wever the same increase in bla ck poverty brings .

a substantia! increase in th e concentration of boa rded-up dw ellings from 2 . 3 % to 3.3%, putting the level of a ba nd o nment within the range of ,

neighborhood instability. In a seg regat ed city, any exogenous econo mic shock that raises the black poverty rate is likely to pu sh black neighb o r­

hoods across the tipp i ng point to begin the process of disinvestment and

abandonment. The same shock delivered to blacks in an integrated city

would leave all neighborhoods stable and economically healthy.

The empiric al correlation upon which this s i mu la t i on . rests was esti­

mated us ing data from sixty metropolitan areas, many of which only recently attained metropolitan rank and therefore have new housing

stocks where the incidence of boarded-up dwellings is near zero. The

simulation has also been c arr i ed out based on correlations estimated

from the city of Philadelphia, where abandonment is conSiderably more

common. 32 Under conditi o ns of complete racial segregation, this alter­

nate simulation predicts that the percentage of board ed up dwellings will

reach 1 0% after the shift i n black incomes, well in excess of the thresho l d

of stability.

Once a process of housing aba ndonment has taken root in one pan of

a community, moreover, it tends to spread outward into ot herwise sta bl e

an d he a l t hy neighborhoods through a process of d iffusion . One of the

principal

me ch anis ms by w h i ch

abandonment

spreads

laterally is

thro u gh fire. n The presence of abandoned a nd boa r d ed up buildings is ·

stro ngly associated with a rson and a high i nci d e n ce of residential fires.14 Once a few b u i l d ing s have burned, nei gh b o rhoo d s no l onger attract sta­

ble h o usehol ds and become a magnet for other social problems, such as rats, litter, drugs, crime, a n d delinquency. 3 5

The lateral spre ad o f d ecay occurs because fires in abandoned building s

inevitably sp re ad to adjacent s tructures that are inhabited, which dis­

places poor families from their dwellings and causes them to seek shelter

in surrou nd ing neighborhoods. The influx of desti t ute and SOcially disor­

gan i zed fire refugees into th ese n eighborho ods increases their hou s i ng densities, levels of homelessness, and dilapidation, setting off new cycles of disinvestment and decline.36 Roderick W alla ce has called this process "urban desertification," d enoting a sequence of events where b y e ntire neig hborh oods b ecome d epop ulated and devoid of stable social institu

-

The Creation of Underdass Communities (io ns.

135

3 7 B y concentrating abandonment, segregation promotes this de­

sertification and builds it in to the residential structure of black commu­ niti e s Th is process of neighborhood decay also occu rs throu gh the loss or withdrawal of commercial institutions. According to Loic Wacquant and .

William Julius Wilson, Chicago's West S ide ghetto lost 7 5 % of its busi­

ness establishments betw e en 1 9 60 and 1 970, and by the 1 980s its North

Lawndale neighborhood incl ude d '.'48 state lottery agents, 50 currency

exch anges, a nd 99 licen sed bars and liquor stores, but only one bank and one supermarket for a popu lati on of some 5 0,000. , , 38 Segregation plays a key role in d ep riv i ng poor b l a ck families of access

goods and services because it interacts with poverty to create neighbor­ hood conditions that make it nearly impo s s ib l e to sustain a v iab le retai l sector. Just as it concentrates poverty during times of economic upheaval, segregation al s o concentrates the loss in income and consumer demand to

that accompany any economic downturn. As a result of racial segrega­

tion, the re fo re, poor blacks live in neighborhoods that typically contain only the barest rudiments of retail trade. They are left without goods and services that are routinely available to me p oo r of other groups.

A simulation ag ai n illustrates the potent ia l of seg reg ati on to concen­

trate income loss and undermine black businesses. As before, we generate

poverty concentra tio ns [rom different combinations of segregation a.nd

bl ack poveny rates, but this time we use the neighborhood poverty rates · to predict me di a n incomes in areas i nh abi ted b y poor blacks.39 In a ra­

cially integrate d city, a shift in the rate of black poverty from 20% to 30% i ncre a s e s the degree o f poverty concentration experienced by poor bl acks and given the empirical association between a neighborhood's poverty rate and median income, this shift is predicted to lower the median neighborhood income for poor blacks from $ 1 3 ,020 before the ,

shift to $ 1 1 ,2 3 5 afterward, yielding a loss of $ 1 ,7 8 5 (see Table 5 . 3 ) .

Assuming there are ro u ghly 3 ,000 hou seho l d s per neighborhood and

th at the median and me an i ncome are about the same, 40 this drop in

i ncome implies a loss of about $ 5 .4 m illio n per neighborhood where

poor blacks live (i.e., 3 ,000 x $ 1 . 78 5 ) . With i n these areas, therefore,

retail p rofi ts will fa ll, services will be cut, and some businesses will fail,

but the drop in income is not enough to bring about the wholesale

elimination of goods and services to poor blacks. The demand density remains fairly high at $ 3 3 . 7 million ( $ 1 1 ,2 3 5 x 3 ,000 ) , and because poor whites and blacks share the same residential areas in an i ntegra ted

A M h R l t.; A N

1 36

APA R T H E I D

city. they e xperien ce these reducrions in retail goods

and services equally

and there is no basis for the fonnation of a racially distinctive u ndercla ss

-::;

'

community.

.. Racial segregation d ramatical l y alters the picture. When the black pov � erty rate rises in a totally segregated city. the drop in i ncome and potential

demand is confined to neighborhoods inhabited exclusively by blacks

,

and primarily by poor blacks. le av ing ihe latter trapped i n neighborhoods

with insufficient in come to sustain a viable retail sector. U nder con d i tions

of complete ra ci al segregation. poor black neighborhoods face a precari- . ous situation

before

the shift in black incomes. with a predicted median

h o u sehol d income of only $8 . 1 60; but after the shift. the predicted me ­ dian income p l u mmetS to $4. 523. for a loss of $ 3 . 6 3 7 (see Table 5 . 3 ) .

In a segregated city. therefore. neighborhoods inhabited by poor blacks

have a weak p otential demand of only $24.5 million before the assumed '

shift in black incomes ( $8 . 1 60 x 3 .000 ) . co mpared with a figure of $ 3 9 . 1

million i n a n integrated city ( $ 1 3 .020 x 3 , 0 0 0 ) , and are thus i n a poor position to ab sorb an economic shock; but racial segregation intensifies

the loss of income created by th e general increase in black poverty and

focuses it precisely upon these very fra gi l e neighborhoo d s After the shift .

in black poverty rates. the potential demand in poor black neighborho ods is almost halved to $ 1 3 .6 million ($4, 5 2 3 x 3 .000 ) , fo r a los