categorically unequal figure 1.1 the stereotype content model

feel toward them is envy. This quadrant embraces the classic mid-dleman minorities, such as Jews in medieval Europe, Chinese inMalaysia, Tutsi in Rwanda, and Indians in East Africa. In a stablesocial structure, people show public respect for and defer to mem-bers of envied out-groups, but if the social order breaks down,these out-groups may become targets of communal hatred and vio-lence because they are not liked and are not perceived as people“like us.”

The top-left quadrant includes out-groups that are viewed aswarm, and thus likable, but as not competent. Those falling intothis category include people who have experienced some misfor-tune but are otherwise perceived as “people like me,” such as thedisabled, the elderly, the blind, or the mentally retarded. One couldimagine being in their shoes but for an accident of fate, and so therelevant emotion is pity. We like the members of these out-groups,but recognizing their lack of competence, we also feel sorry for

Categorically Unequal

12

PitiedOut-group

EsteemedIn-group

EnviedOut-group

DespisedOut-group

Competence

Warmth

Source: Author’s compilation.

Figure 1.1 The Stereotype Content Model

Roosevelt and his New Deal and then under Harry S. Truman andhis Fair Deal.

The economic policies of both “Deals,” in combination withstrong deflationary pressures during the 1930s and the SecondWorld War, led to a remarkable compression of the income distribu-tion and a sharp decline of inequality during the 1940s (Goldin andMargo 1992). By 1952, when Republicans finally reoccupied theWhite House, the share of income going to the top 10 percent of tax-payers had fallen from 45 percent to 32 percent; that going to thetop 5 percent had dropped from 35 percent to 21 percent; and thataccruing to the top 1 percent had declined from 20 percent to 10percent. These figures represent respective declines of roughly 30,40, and 50 percent from the peak levels of inequality in 1928 mea-sured by Piketty and Saez.


34

50

45

40

35

30

25

20

15

10

5

0

Perc

enta

ge o

f Inc

ome

1922

1927

1932

1937

1942

1947

1952

1957

1962

1967

1972

1977

1982

1987

1992

1997

2002

Year

NeoliberalRestructuringPostwar Consensus

New Deal–Fair Deal

Laissez-Faire Era

Top 10%

Top 5%

Top 1%

Figure 2.1 Share of Income Earned by Top Segments ofTaxpayer Income Distribution, 1922 to 2002

Source: Piketty and Saez (2003).

available until the 1990s, they are not considered here.) Althoughwhite and black households both experienced rising incomes overthe period, very little progress was made in closing the racial gap,despite policies and programs instituted during and after the civilrights era. In real terms, the size of the racial gap actually grew from1972 to 1988, going from $16,700 to $19,700 before falling back to$18,000, around which it has fluctuated ever since. In other words,three decades of affirmative action efforts and antidiscriminationpolicies have not closed the gap between black and white incomes.To put it bluntly, racial inequality remains a basic feature of the U.S.stratification system.

Since the civil rights era, of course, the United States has changeddemographically in rather dramatic ways, and Latinos recentlyovertook African Americans as the nation’s largest minority. His-


38

$50,000

$45,000

$40,000

$35,000

$30,000

$25,000

$20,000

Inco

me

(Con

stan

t 200

2 D

olla

rs)

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

Year

Whites

Latinos

Blacks

Figure 2.2 Median Household Income, 1967 to 2003

Source: U.S. Bureau of the Census.

households after the great U-turn are clearly indicated in the figure.Whereas the incomes of households at the bottom of the distribu-tion stagnated and the median inched slowly upward, incomes atthe top of the distribution rose dramatically. Among households inthe lowest quintile, for example, average income stood at $8,800 in1972 and had risen to $10,200 by 2002, an increase of $1,400, or 16percent, in twenty years. Those at the middle of the income distri-bution fared little better. Over three decades, the median householdincome grew from $37,000 to $43,400, for a real gain of $6,400, or 17percent. In contrast, average income for the highest quintile in-creased by 66 percent, going from $88,700 in 1972 to $147,000 in2002, for a real gain of $58,300.

Figure 2.3 illustrates a basic feature of inequality in the post-in-


40

$300,000

$275,000

$250,000

$225,000

$200,000

$175,000

$150,000

$125,000

$100,000

$75,000

$50,000

$25,000

$0

Mea

n In

com

e (C

onst

ant 2

002

Dol

lars

)

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

Year

Top 5%

Top 20%

Bottom 20%

Median

Figure 2.3 Average Constant Income Earned by Households atDifferent Points in the Income Distribution, 1967 to2002


than from a deterioration in the economic position of men. In 1973the median personal income for males stood at $28,900; thirty yearslater it was only $29,900, a real increase of only $1,000, or 3.5 per-cent. Among women, median personal income rose from $10,000 in1973 to $17,300 in 2003, an increase of $7,300, or 73 percent. Al-though this pay raise is nothing to scoff at, thirty years of progressstill leaves the average woman receiving less income than the aver-age male got back in 1951, and 43 percent less than that of the aver-age male today.

Earlier we noted that median household income rose by 17 per-cent from 1972 to 2002. Given that male median income rose byonly 3.5 percent over the same period, there is only one reasonableexplanation for the continued increase in household income during

Rise and Fall of Egalitarian Capitalism

43

$35,000

$30,000

$25,000

$20,000

$15,000

$10,000

$5,000

$0

Inco

me

(Con

stan

t 200

2 D

olla

rs)

1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002

Year

Males

Females

Figure 2.4 Median Personal Income Earned by Males andFemales, 1947 to 2002


a period of stagnant male income: female household members en-tered the labor force in large numbers to generate additional in-come that offset the declining fortunes of men. Although middle-income households managed to achieve modest income gains afterthe great U-turn, more household members—usually wives—hadto work to do so. In other words, in the post-industrial politicaleconomy of the United States, it increasingly requires two earnersfor a household to remain in the middle class.

The importance of having two incomes is illustrated in figure 2.6,which shows the change in median income since 1987 among one-and two-earner households. Naturally, two-earner householdshave more income than those with one breadwinner, but we are in-terested in relative change, not absolute amounts, so each series hasbeen divided by its value in 1987 to indicate proportional growth


44

0.6

0.55

0.5

0.45

0.4

0.35

0.3

Rat

io

1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002

Year

Figure 2.5 Female Personal Income as a Proportion of MalePersonal Income, 1947 to 2002


since that date. As can be seen, from 1987 through 1990 the incomesof one- and two-earner households moved together, but thereafterthey began to depart. The recession of the early 1990s hit all house-holds hard, but whereas two-earner households saw their incomesdrop, they never dipped below the median for 1987, and by 1992they had begun to rebound. In contrast, the median income of one-earner households fell to 96 percent of the 1987 value and did notmake it back to parity until 1997. Thus, one-earner households ex-perienced an entire decade with no growth in income, whereastwo-earner households were able to hang on during the recessionand increase their incomes by 10 percent by 1997. By the year 2003,two-earner households enjoyed a median income that was 19 per-cent above what they reported in 1987, whereas the median incomeearned by one-worker households was only 8 percent more.


45

1.2

1.15

1.1

1.05

1

0.95

Val

ue R

elat

ive

to 1

987

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

Year

Two-Earner Households

One-EarnerHouseholds

Figure 2.6 Incomes of One- and Two-Earner HouseholdsRelative to Their Values in 1987


America’s New InequalityThe purpose of this chapter has been to provide a general descrip-tion of what happened to Americans with respect to income strati-fication during the twentieth century. As we have noted, the pastone hundred years have been tumultuous for both the world andthe United States, including two periods of globalization, both ac-companied by massive shifts in trade, investment, population, andsocial structure. The first era of globalization was accompanied byhigh levels of inequality that in the United States peaked in the late1920s, when the effects of the First World War, isolationism, andprotectionism brought national and global economies crashingdown.


47

1.4

1.3

1.2

1.1

1

0.9

0.8

Val

ue R

elat

ive

to 1

973

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

Year

Peak Peak Peak

FemaleParticipation

MalePersonalIncome

Female Ratein 2003: 59.3

FemaleRate in 1973: 42.0

Figure 2.7 Trends in Female Labor Force Participation andMale Personal Income


doubled their private recruitment efforts (Durand and Arias 2000).The number of contract workers entering the United States grew to17,000 in 1920, and legal immigration reached an unprecedented51,000 persons. American insecurities about all things foreign cameto a head during the recession that followed the war, and the redscare of 1918 to 1921 was accompanied by a wave of anti-immigranthysteria. Congress passed the first quota law in 1921 to curtail im-migration from southern and eastern Europe and enacted an evenstricter version in 1924. During this period, both legal immigrationand contract migration by Mexicans fell, reaching lows in 1922 of18,000 and 12,000, respectively. But economic recovery led to thesustained economic boom known as the Roaring Twenties, andgiven the new restrictions on European immigration and steadily

Building a Better Underclass

121

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0

Num

ber

of M

igra

nts

1900

1902

1904

1906

1908

1910

1912

1914

1916

1918

1920

1922

1924

1926

1928

1930

Year

Gentlemen'sAgreement

MexicanRevolution

FirstQuota Act

Legal Immigrants Apprehensions Temporary Workers

U.S. EntersWorld War I Founding of

Border Patrol

Figure 4.1 Mexican Migration to the United States, 1900 to 1930

Source: U.S. Immigration and Naturalization Service.

Table 4.1 Distribution of People Depicted on Magazine Covers, by Race, Ethnicity, andGender, Compared to Distribution Among Immigrants and U.S. Population

CoverActual Data Cover Photos Illustrations

Variable and Immigrants, U.S. Total, Non- Non-Distribution 1990s 2000 Immigrants immigrants Immigrants immigrantsNational origin, race, ethnicity

European/white 15% 69% 10% 71% 0% 69%Asian 31 4 40 12 9 0Afro-Caribbean/black 15 13 23 2 46 2Latin American/Hispanic 37 13 26 1 45 25

GenderMale 55 51 79 70 73 100

Sources: Office of Immigration Statistics, U.S. Department of Homeland Security; U.S. Bureau ofthe Census; Chavez (2001).

sented racial inferiors rising above their assigned station, and in thecharged atmosphere of wartime Los Angeles, anti-Mexican riotingbroke out. On June 3, 1943, a group of servicemen on leave com-plained that they had been assaulted by a gang of pachucos wear-ing zoot suits (Obregón Pagán 2003). In response, an angry mob ofwhite soldiers and civilians headed into the Mexican barrio of EastLos Angeles, where they attacked all males wearing zoot suits,beating them severely while ripping off the offensive garments andburning them on the spot.

Rather than protecting U.S. citizens of Mexican origin, the LosAngeles police swept into the barrio and arrested hundreds of al-


125


1,200,000

1,000,000

800,000

600,000

400,000

200,000

0

Num

ber

1940

1942

1944

1946

1948

1950

1952

1954

1956

1958

1960

1962

1964

Year

World War II Ends

Bracero ProgramBegins

OperationWetback

Legal Immigrants Apprehensions Temporary Workers

Bracero ProgramEnds

Source: U.S. Immigration and Naturalization Service.

braceros or legal immigrants during the late 1950s, the total inflow ofMexicans through legal channels fell to just 62,000 per year from1965 to 1985—13 percent of its former level. The gap between the de-mand for visas on the part of employers and workers and the paltrynumber offered by the government was increasingly made upthrough undocumented migration, and annual apprehensions alongthe border climbed steadily from 55,000 in 1965 to 1.6 million in 1985.

In essence, the shift in U.S. immigration policy after 1965 trans-formed Mexican migration from a de jure guest-worker programbased on the circulation of braceros into a de facto guest-workerprogram based on the circulation of undocumented migrants (Du-


129


1,800,000

1,600,000

1,400,000

1,200,000

1,000,000

800,000

600,000

400,000

200,000

0

Num

ber

1965

1968

1971

1974

1977

1980

1983

1986

1989

1992

1995

1998

2001

2004

Year

HemisphericQuota TakesEffect

CountryQuotasImposed

Limits to FamilyImmigration Imposed

Legal ImmigrantsApprehensions

Temporary Workers

Deportations

ImmigrationReform and Control Act

Sources: U.S. Immigration and Naturalization Service; U.S. Department ofHomeland Security.

Thus, U.S. immigration and border policies transformed Mexi-can immigration from a circular flow of male workers into a settledpopulation of families. By the year 2005, the total number of undoc-umented Mexicans present north of the border had reached 6.1 mil-lion, and the total number of foreign-born Mexicans had climbed to11 million. These figures imply that one of every ten people born inMexico now lives in the United States, and 55 percent are present inillegal status. At the same time, immigrants now constitute a grow-ing fraction of all Mexican Americans, roughly 40 percent as of2005, and the large number of undocumented among them meansthat midway in the first decade of the twenty-first century, morethan half of all Mexican-born persons—and more than one-fifth ofall persons of Mexican origin—lack any social, political, or eco-


142

Figure 4.4 Growth of the Undocumented Mexican Population ofthe United States, 1965 to 2005

7,000

6,000

5,000

4,000

3,000

2,000

1,000

0

Num

ber

in T

hous

and

s

1965

1968

1971

1974

1977

1980

1983

1986

1989

1992

1995

1998

2001

2004

Year

Sources: Woodrow-Lafield (1998); Hoeffer, Rytina, and Campbell (2006).

groups: South Americans, Latinos, Mexicans, farmworkers, andAfricans. The most despised immigrant group of all is undocu-mented migrants, who rate lowest of all on both warmth and com-petence, such that they approach the portion of the graph usuallyreserved only for the most detested and socially stigmatizedgroups, such as criminals and drug dealers. In societal terms, this isdangerous territory, since it implies that undocumented migrantsare not perceived as fully human at the most fundamental neurallevel of cognition, thus opening a door to the harshest, most ex-ploitive, and cruelest treatment that human beings are capable ofinflicting on one another.

Categorical Stratification After 9/11Just as the criminal justice system has emerged as a new and pow-erful institution in promoting categorical inequality between blacks


150

War

mth

high

high

low

low

Competence

Homeless People

Housewives

Elderly People

Italian

Irish

JapaneseKorean

Chinese

RichPeople

AsianProfessionals

Tech Industry

Canadian

CollegeStudents

European

Indian

ThirdGeneration

DocumentedAmericansMexican

Latino

Poor People

African

SouthAmerican

Farmworker

UndocumentedGerman

RussianFrench

FirstGeneration

EasternEuropean

Vietnamese

MiddleEastern

Figure 4.5 Stereotype Content Model Applied to Immigrants

Source: Lee and Fiske (2006). Reprinted with permission from Elsevier.

onward, Hispanics replaced blacks at the bottom of the male earn-ings hierarchy.

The top two lines show trends in the relative earnings of His-panic and black women to reveal a similar switching of positions,one that occurred even earlier than among minority men. In 1972Hispanic women earned the same income as white women, whileblack women earned 92 percent of what their white counterpartsearned. During the 1970s and 1980s, however, the earnings ofwomen in both minority groups deteriorated relative to those ofwhite women, but the decline was more rapid among Hispanicwomen. In 1981 the two lines crossed, and from then on Hispanicwomen replaced black women at the bottom of the female earningshierarchy.


152

1

0.9

0.8

0.7

0.6

0.5

Inco

me

Rat

io

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

Year

Immigration Reformand Control Act

Hispanic Men

Black Men

Hispanic Women

Black WomenOperationBlockade

Figure 4.6 Hispanic and Black Personal Income as a Ratio ofWhite Income


Despite this early crossover, however, the Hispanic-white andblack-white income ratios remained quite close to one another untilIRCA passed in 1986. After this date, the deterioration in black fe-male income slowed down, and then in the early 1990s it reversedand began to move upward. In contrast, the deterioration in His-panic women’s earnings accelerated, and the two income ratios be-gan to pull apart at a rapid pace, yielding a widening gap. From 1987to 2002, the ratio of black female income to white female income rosefrom .80 to .96, while the income ratio for Hispanic females fell from.80 to .70 before coming back up to end the decade at .77.

The shifting fortunes of Hispanics and African Americans in theU.S. labor market are also clearly reflected in American poverty sta-tistics. Historically, rates of Hispanic poverty were far below thoseof blacks, but over the course of the 1980s and 1990s the differential


153

1.2

1.1

1

0.9

0.8

0.7

0.6

Rat

io to

Whi

te P

over

ty

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

Year

Immigration Reformand Control Act

OperationBlockade

Figure 4.7 Ratio of Hispanic to Black Poverty Rates


Relations Act in 1935, whereupon union membership surged, morethan tripling from around 10 percent of workers in 1935 to one-third of all workers just ten years later.

With the passage of the Taft-Hartley Act in 1949, however, unionmembership ceased growing and then began to decline after 1959when Landrum-Griffin further strengthened the hand of employ-ers. From a peak rate of 33 percent in 1954, unionization fell to 23percent by 1980, at which time Ronald Reagan was elected to thepresidency. His first year in office proved to be pivotal. In 1981 theunion representing air traffic controllers voted to go on strike forbetter working conditions and improved safety requirements. In-voking Taft-Hartley, President Reagan declared the strike illegaland promised to fire all those who did not return to work within

Remaking the Political Economy

163

35.0

30.0

25.0

20.0

15.0

10.0

5.0

0.0

Uni

on M

embe

rs/N

onfa

rm W

orke

rs

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Year

Air TrafficControllers'Strike

Landrum-Griffin Act

National LaborRelations Act

Taft-Hartley Act

Figure 5.1 Level of Unionization in the United States, 1900 to2004

Source: Carter et al. (2006).

share of workers enjoying this earnings premium declines as well.In addition, with unionization no longer viewed as a credible threatby employers, they have little incentive to pay workers higherwages as an inducement not to form a union. Studies reveal that inthe past the threat of unionization served to boost the wages of allworkers, even those who were not union members (Corneo andLucifora 1997; Freeman and Medoff 1984; Leicht 1989; Neumarkand Wachter 1995), including even midlevel managers (Rosenfeld,forthcoming).

As a result, over time there has been a strong positive relation-ship in the United States between the frequency of strikes and the


165

1.2

1

0.8

0.6

0.4

0.2

0

Freq

uenc

y R

elat

ive

to 1

946

1936

1940

1944

1948

1952

1956

1960

1964

1968

1972

1976

1980

1984

1988

1992

1996

2000

Year

SuccessfulElection

Strikes

Figure 5.2 Relative Frequency of Strikes and Successful UnionElections, 1936 to 2000

Source: Carter et al. (2006).Note: 1946 = 1.

figure 5.4 illustrates the effect of these wage reductions on the wel-fare of families supported by low-wage workers.1 Specifically, itcomputes how much income would be earned working full-time(38.5 hours per week) for a full calendar year (52 weeks) and ex-pressing the result as a share of the official poverty level for a fam-ily of four. As can be seen, never in the course of the past fifty yearshas the official minimum wage been sufficient to lift a family withone breadwinner out of poverty, although during the height ofJohnson’s Great Society it came close. In 1968 a minimum-wageworker employed full-time generated annual earnings equal to 93percent of the federal poverty limit.

The figure has been labeled to show Republican versus Democ-ratic administrations. After the 1960s, Republican administrations


167

Figure 5.3 Real Value of the U.S. Minimum Wage (2002Dollars), 1938 to 2004

10

9

8

7

6

5

4

3

Wag

e pe

r H

our

1938

1943

1948

1953

1958

1963

1968

1973

1978

1983

1988

1993

1998

2003

Year

Source: U.S. Bureau of Labor Statistics.

were very clearly associated with falling minimum wages, whereasDemocratic administrations were associated with steady, fluctuat-ing, or rising minimum wages. A well-controlled and more com-prehensive analysis by Larry Bartels (2004) showed that Republi-can presidents indeed produce greater income growth for rich thanpoor families and that rising inequality is reliably associated overtime with Republican administrations. By minimizing the mini-mum wage, Congress by the year 2005 guaranteed that the poorestworkers in America would make no more than 53 percent of thefederal poverty rate even if they worked full-time. In other words,even two full-time workers employed at the minimum wage wouldearn barely enough money to keep a family of four above thepoverty threshold of $19,800 per year.


168

0.95

0.9

0.85

0.8

0.75

0.7

0.65

0.6

0.55

0.5

Prop

orti

on o

f Pov

erty

Ear

ning

s

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Year

Republican Republican Republican

Rep

ublic

anDemocrat

Dem

ocra

t Democrat

Figure 5.4 Earnings for a Full-Time Minimum-Wage Job as aShare of Poverty Earnings for a Family of Four

Sources: U.S. Bureau of Labor Statistics; U.S. Bureau of the Census,Statistical Abstract of the United States.

From a per capita level of around 12 cents in 1960, spending on an-tipoverty programs more than tripled to peak at 30 cents in 1972 be-fore falling precipitously during Nixon’s second term. The free fallin poverty-related social services was checked early in the Carteradministration but then resumed during the inflation of the late1970s as Congress failed to make adjustments to transfers to reflectthe rising cost of living. Throughout the presidencies of RonaldReagan and George H. W. Bush, antipoverty spending remainedflat at only slightly above the level it had been on the eve of theGreat Society, a level that persisted through the early years of theClinton administration.

Thus, in terms of social spending on the poor, by the early 1990sit was as if Johnson’s war on poverty had never happened. Even


170

14,000

12,000

10,000

8,000

6,000

4,000

2,000

0

2003

Dol

lars

per

Une

mpl

oyed

Wor

ker

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

Year

Nixon-Ford Carter Reagan-Bush Clinton

Figure 5.5 Spending on Unemployment Benefits perUnemployed Worker

Source: U.S. Bureau of the Census, Statistical Abstract of the UnitedStates.

this reduced level of spending on the poor was too much, however,for congressional conservatives, who in 1996 succeeded in passingthe Personal Responsibility and Work Opportunity ReconciliationAct (PRWORA), essentially forcing mothers of dependent childrenoff of welfare rolls and into the labor force by ending the entitle-ment to open-ended income transfers. Aid to Families with Depen-dent Children (AFDC) was replaced by a new program called Tem-porary Assistance for Needy Families (TANF), which gave mothersonly two years of assistance before requiring them to work. Accord-ing to the terms of the legislation, 25 percent of all families weresupposed to leave the welfare rolls by 1997, with 50 percent ex-pected to be working by 2002.

As indicated by figure 5.7, which depicts the percentage of poor


171

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0

2003

Dol

lars

per

Per

son

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

Year

Roosevelt-Truman

Eis

enho

wer Kennedy-

JohnsonNixon-

Ford

Car

ter Reagan-Bush

Figure 5.6 Per Capita Public Spending on Social Services for thePoor, 1930 to 1995


families receiving AFDC or TANF benefits by year, the effect of thenew law on welfare use was sudden and dramatic. Historically therate of welfare use in the United States had been low because Con-gress wrote the Social Security Act to exclude black workers anddelegate to states the right to determine eligibility. Given the over-representation of African Americans among the poor and the dis-proportionate location of all poor families in southern states withstringent eligibility requirements and low payments, only a tinyfraction of the poor actually received welfare before the 1960s. In-deed, until 1965 the share of poor families receiving income trans-fers ranged only from 10 to15 percent.

The passage of civil rights laws and other federal reforms duringthe Great Society made it more difficult for states to deny relief to


172

Figure 5.7 Poor Families Receiving AFDC or TANF, 1960 to 2005

Sources: Carter et al. (2006); U.S. Bureau of the Census.

50

45

40

35

30

25

20

15

10

5

0

Perc

enta

ge o

n W

elfa

re

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Year

Bankruptcy of NWRO

Founding of NWRO

PRWORA

into legislation. The 1996 Personal Responsibility and Work Oppor-tunity Reconciliation Act was a self-proclaimed effort to “end wel-fare as we know it,” and as is obvious from figure 5.7, the effect onusage rates was immediate and profound. From a figure of 36 per-cent in 1996, the share of poor families receiving federal incometransfers fell sharply to just 10 percent in 2004, a level of welfare uti-lization last observed in 1963. Along with the decline in welfare re-ceipt came a drop in spending for other entitlements that were tiedto it, such as food stamps. As shown in figure 5.8, real expendituresfor food stamps fell dramatically as the new welfare policy pushedwomen off of welfare and into ineligibility for food stamps.Whereas federal spending on food stamps had totaled $840 perpoor person on the eve of welfare reform, by 2000 the figure had


174

Figure 5.8 Real per Capita Spending on Food Stamps, 1970 to2002

Source: U.S. Bureau of the Census, Statistical Abstract of the UnitedStates.

900

800

700

600

500

400

300

200

2003

Dol

lars

per

Poo

r Pe

rson

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

Year

Nixon-Ford

Carter Reagan-Bush Clinton

Nixon and Ford administrations and stabilized at 8.8 workers perthousand under President Jimmy Carter.

The decline in federal employment resumed during the first Rea-gan administration, but then recovered somewhat before experi-encing a sustained decline over the course of the two Clinton administrations. The election of George Bush in 2000, however,brought an unprecedented assault on federal employment: thenumber of federal workers per thousand fell to a record low of 6.2within two years of his assuming office, the fastest drop in federalemployment in American history. In relative terms, the federal gov-ernment is now smaller than at any point since 1940, reflecting itsreduced capacity for managing and regulating the U.S. politicaleconomy.


176

Figure 5.9 Nondefense Government Workers per 1,000Population

Sources: Carter et al. (2006); U.S. Bureau of the Census, StatisticalAbstract of the United States.

10

9.5

9

8.5

8

7.5

7

6.5

6

Num

ber

per

1,00

0 Pe

rson

s

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Year

ClintonKennedy-Johnson

Nixon-Ford

Carter Reagan-Bush Bush

5.11 plots the average interest rate charged on credit cards com-pared with the interest rate prevailing in the rest of the economy. Ascan be seen, the gap between the two series was relatively narrowduring the 1970s, averaging only 4.4 percentage points. Over thecourse of the 1980s and 1990s, however, the gap steadily widened,earning credit card lenders, in real terms, more money per dollarloaned. During the period 2000 to 2005, the gap between the inter-est rate on credit cards and the prevailing rate averaged 8.2 percent-age points.

The rising mountain of consumer debt is not without risks tobanks, of course, despite high interest rates and record profits, for ifthe debt becomes too burdensome to sustain, some credit cardholders simply walk away from their obligations by declaringbankruptcy. Historically, bankruptcy laws in the United States were


178

Figure 5.10 Per Capita Consumer Debt, 1945 to 2005

Source: U.S. Federal Reserve.

8

7

6

5

4

3

2

1

0

2006

Dol

lars

per

Per

son

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

2005

Year

relatively lenient with consumers who borrowed over their heads,allowing them to go to court and wipe away most of their obliga-tions to begin their financial lives afresh. For bankers left with theunredeemable loans, however, this act represented an “abuse” ofthe bankruptcy system, and at their behest, Congress in 2005passed the Bankruptcy Abuse Prevention and Consumer Protec-tion Act.

The new law placed barriers on the road to bankruptcy andforced struggling consumers to pay down more of their accumu-lated debt. After 2005, debtors were obligated to undergo “creditcounseling” in a government-approved program before being al-lowed even to file for bankruptcy, and then once a filing had beenmade, they had to pay all back taxes before being allowed to pro-


179

Figure 5.11 Interest on Credit Card Debt Versus Overall InterestRates, 1975 to 2005

Source: U.S. Federal Reserve.

20

18

16

14

12

10

8

6

4

2

0

Inte

rest

Rat

e

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

Year

Credit Card Interest

Overall Interest

percent in 1917, and reaching 77 percent in 1918, again with no dif-ferentiation between capital gains and regular income. The end ofthe war brought a slight reduction in taxes, with the top rate fallingto 73 percent until 1922, when a Republican administration andCongress lowered the top rate to 58 percent before easing it steadilydownward to 44 percent in 1923 and 25 percent in 1924, where it re-mained until the election of Franklin Roosevelt in 1932. At the sametime, beginning in 1922, Congress also differentiated capital gainsfrom regular income, reducing the rate of tax on gains from the saleof assets such as stocks and bonds to just 12.5 percent.

The advent of the New Deal revolutionized American tax policyby introducing a progressive system that taxed additional incomeat steadily higher rates as earnings rose. From 1932 through 1935,the wealthiest Americans paid 63 percent on regular income over$1 million, and the tax on capital gains rose from 12.5 percent to


181

Figure 5.12 Top Marginal Tax Rates and Capital Gains TaxRates, 1913 to 2005

Sources: Carter et al. (2006); U.S. Bureau of the Census, StatisticalAbstract of the United States.

100

90

80

70

60

50

40

30

20

10

0

Tax

Rat

e

1910

1915

1920

1925

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

2005

Year

Rate on Top Income

Rate on Capital Gains

ity to the system, raising the average tax rate to nearly 40 percentfor wealthy households while leaving the rates paid by affluent andpoor household about the same.

Most recently, George W. Bush’s tax cuts reduced the averagerate for all three classes, but progressivity was substantially re-duced compared to the Clinton years, and the burden of paying forgovernment was shifted downward in the income distribution. Asof 2005, poor households paid 20 percent of their income in taxes,affluent households paid 27 percent, and wealthy households paid34 percent—not quite the bonanza for the rich achieved at the endof the second Reagan administration, but a far cry from the pro-gressive taxation that prevailed from 1932 to 1980 under arrange-ments set by the New Deal.


184

Figure 5.13 Average Tax Rates for Households Earning $20,000,$200,000, and $2 Million per Year

Source: Author’s calculations.

80

70

60

50

40

30

20

10

0

Tax

Rat

e

1965 1970 1975 1980 1985 1990 1995 2000 2005

Year

ClintonNixon-Ford Reagan-Bush Bush

$20,000 Income

$2 Million Income

$200,000 Income

represents a complex choreography, a “dance of ideology and un-equal riches” in which polarization inhibits support for redistribu-tion, which produces greater inequality, which feeds more polariza-tion, which generates even more inequality.

They argue that the power of this choreography has been exacer-bated by an increase in the number of poor people without politicalrights who cannot vote and thus cannot exert political pressurefrom the bottom in support of redistributive policies. They point tothe post-1965 revival of immigration and note that the correlationbetween the percentage foreign-born and political polarization is0.92. The current potential for polarization and inequality is evengreater than they suppose, however, for they do not take into ac-count the huge increase in the share of immigrants who are not


189

Figure 5.14 Indices of Income Inequality and PoliticalPolarization Developed by McCarty, Poole, andRosenthal (2006)

Source: Piketty and Saez (2001), figure 1.2. Reprinted with permis-sion from MIT Press.

21

19

17

15

13

11

9

7

.85

.8

.75

.7

.65

.6

.55

.5

Perc

enta

ge S

hare

Pola

riza

tion

Ind

ex

1913

1917

1921

1925

1929

1933

1937

1941

1945

1949

1953

1957

1961

1965

1969

1973

1977

1981

1985

1989

1993

1997

Year

r = .78

Income Share of Top 1% House Polarization Index

tial concentration of advantage and deprivation is best measuredusing the P* isolation index, which yields the proportion of poor oraffluent families in the neighborhood of the average poor or afflu-ent person (Massey and Denton 1988; Massey and Eggers 1990).Figure 5.16 shows trends in the concentration of poverty using P*indices computed for 1970 through 2000, and figure 5.17 does thesame for the concentration of affluence.

At all points in time over the past thirty years, affluence has beenmore concentrated spatially than poverty. In other words, thosewith money are more likely to live in homogenously affluent envi-ronments than are those without money to live in homogenouslypoor environments. Geographic concentration increased at both


193

Figure 5.15 Income Segregation in Large Metropolitan Areasof the United States, 1970 to 2000

Source: Massey and Fischer (2003).

0.45

0.4

0.35

0.3

0.25

0.2

0.15

0.1

Aff

luen

t-Po

or D

issi

mila

rity

1970 1980

0.398

0.287

0.43

0.373

1990 2000

Year

194

Figure 5.16 Concentration of Poverty in Large U.S.Metropolitan Areas, 1970 to 2000


0.3

0.28

0.26

0.24

0.22

0.2

0.18

0.16

0.14

0.12

0.1

Pove

rty

Isol

atio

n In

dex

1970 1980

0.162

0.136

0.279

0.246

1990 2000

Year

0.4

0.35

0.3

0.25

0.2

0.15

0.1

Aff

luen

ce Is

olat

ion

Ind

ex

1970 1980

0.276

0.308

0.3560.338

1990 2000

Year

Figure 5.17 Concentration of Affluence in Large U.S.Metropolitan Areas, 1970 to 2000



competence. Figure 6.1 reproduces Eckes’s results to show how dif-ferent subtypes of women are stereotypically perceived within con-temporary society.

It is immediately apparent that most of the subtypes cluster inthe top left-hand or bottom right-hand quadrants of the socialspace, zones normally reserved for pitied and envied out-groups.No women really fall into the despised quadrant at the bottomright, and none really fall into the esteemed quadrant at the top left.Those closest to the space normally reserved for despised out-groups are “bourgeois women” and “chicks,” and those closest tothe space usually occupied by esteemed in-groups are “societyladies” and “confident types.” But for the most part the zones asso-ciated with esteem and contempt are empty.

Instead, the social spaces associated with pity and envy are


216

4.5

4

3.5

3

2.5

2

1.51.5 2 2.5 3 3.5 4 4.5

Competence

War

mth

Cluster A

Cluster B

Cluster C

Cluster D

Housewife

Wallflower Secretary

Naive Type

ChickBourgeois

Trendy

Society LadyConfident Type

VampIntellectual

HippiePunk

Women's Libber

Career Woman

Feminist

Typical Woman

Figure 6.1 Stereotype Content Model Applied to SelectedFemale Subtypes

Source: Eckes (2002). Reprinted with permission from Springer-VerlagGmbH.

Although the level of occupational segregation between menand women remained fairly constant from 1860 to 1960, the under-lying size and structure of the labor market changed dramatically.Whereas in 1860 the female workforce numbered just 5.2 million,only 16 percent of whom held white-collar jobs, by 1960 the num-ber of female workers stood at 22.2 million, and 53 percent heldwhite-collar jobs (see Carter et al. 2006; figure 6.2). Overall, the la-bor force grew by 144 percent, and the structure of employmentshifted decisively away from agriculture—first toward manufac-turing and then toward services.

The fact that gender segregation remained constant despite thesemassive shifts suggests that a great deal of boundary work was

Engendering Inequality

219

Figure 6.2 Occupational Segregation, Female Labor ForceParticipation, and Female Concentration in White-Collar Jobs, 1860 to 2000


80

70

60

50

40

30

20

10

0

Perc

enta

ge

1860

1870

1880

1890

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

Year

Male-Female Segregation

Percentage White-Collar

Labor Force Participation

ers on the right. The y-axis indicates the percentage of women inthe different occupational categories. As can be seen, relativelymore women work in nonmanual jobs, but within the white-collarsector the presence of women falls as one moves up the ladder ofoccupational earnings and prestige; the same pattern prevails in theblue-collar sector.

The rise of horizontal segregation in the United States is sug-gested by the data in figure 6.4, which plots the percentage of fe-male workers in manual and nonmanual jobs from 1860 through1990. Up through 1900, there was no differentiation between thetwo sectors. Few women were in the labor force, but they were


221

Figure 6.3 Charles-Grusky Model of Gender Segregation

Source: Charles-Grusky (2004).

100.0

80.0

60.0

40.0

20.0

0.0

Perc

enta

ge F

emal

e

Profes

sional

Man

ager

ial

Clerica

l

Serv

ices

Sales

Crafts

Operati

ves

Farm

worker

s

Laborer

s

OccupationManualNon-manual

equally likely to work in the manual or nonmanual sector. There-after, female specialization in white- as opposed to blue-collar jobsincreased at an accelerating rate, producing a widening gap in thegender composition of the two sectors. Whereas 55 percent of non-manual workers were female by 1990, only 31 percent of manualworkers were women.

Figure 6.5 illustrates the rise in vertical gender segregation after1900. Up through 1900, those few females who worked were just aslikely as men to be employed as managers, proprietors, clericalworkers, or sales workers, and females were actually overrepre-sented in professional occupations compared to their small share inthe labor force. After 1900, however, women came increasingly todominate clerical occupations, until by 1990 nearly 80 percent ofclerical workers were female. Women also came to constitute a


222

Figure 6.4 Percentage of Female Workers in Manual andNonmanual Occupations


60.0

50.0

40.0

30.0

20.0

10.0

0.0

Perc

enta

ge F

emal

e

Year18

6018

7018

8018

9019

0019

1019

2019

3019

4019

5019

6019

7019

8019

90

Nonmanual Occupations

Manual Occupations

growing share of sales workers. After 1910, however, the share ofwomen in professional occupations leveled off and then fell gradu-ally through 1960 as the number of female managers and propri-etors grew very slowly and at levels below their rising share of thetotal labor force, a pattern that lasted until 1970.

The most comprehensive account of the gendered transforma-tion of the American labor market in the twentieth century is that ofGoldin (1990). After 1900, industrialization increased the demandfor office workers. Managers, professionals, clerks, secretaries, andsales personnel became increasingly important in creating value forbusiness enterprises. To meet the surge in demand for nonmanualworkers, legislatures passed new laws mandating school atten-dance and increased funding for secondary education. As a conse-quence, levels of high school graduation rose among both malesand females, but with factories providing an attractive and lucra-


223

Source: Historical Statistics of the United States Millennial Edition.

Figure 6.5 Percentage of Female Workers in NonmanualOccupations

80.0

70.0

60.0

50.0

40.0

30.0

20.0

10.0

0.0

Perc

enta

ge F

emal

e

Year

Professionals

1860

1870

1880

1890

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

ManagersProprietorsClerical

Sales

231

Figure 6.6 Mean SEI Score and Percentage with CollegeDegree, by Occupational Status Quintile

100.0

90.0

80.0

70.0

60.0

50.0

40.0

30.0

20.0

10.0

0.0SEI S

core

or

Perc

enta

ge o

f Col

lege

Gra

dua

tes

Status Quintile

College-MaleCollege-Female

Status-MaleStatus-Female

Q1 Q2 Q3 Q4 Q5

Source: Jacobs (2001).

Figure 6.7 Earnings by Occupational Status

$70,000

$60,000

$50,000

$40,000

$30,000

$20,000

$10,000

$0

Ear

ning

s

Status Quintile

Males

Females

Q1 Q2 Q3 Q4 Q5

Source: Jacobs (2001).

Figure 6.8 puts aside education and shows the degree to whichother measured factors statistically account for observed differ-ences between male and female earnings. As already noted, expe-rience accounts for only 16 percent of the gap, and given the weak-ened state of unions, as described in the last chapter, it is hardlysurprising that unionization accounts for only 5 percent of the gap.Likewise, racial composition explains only 4 percent of the gap,suggesting that women earn roughly the same amounts regardlessof race. By far the two most important measured factors account-ing for the male-female earnings gap are industrial location (33percent) and occupation (42 percent), implying that if the distribu-tion of men and women across occupations were somehow equal-ized, up to three-quarters of the earnings differential would disap-pear.


233

Figure 6.8 Contribution to Gender Pay Gap of MeasuredCharacteristics

16%

4%

42%

5%

33%

Experience

Race

Occupation

Industry

Unionization

Source: Blau and Kahn (2006).

collar occupational distribution. There has been little or no move-ment of women into skilled crafts and operative categories, the twomost highly remunerated blue-collar occupations. Moreover, al-though the share of women increased slightly among farmworkers,it rose substantially among unskilled laborers, the lowest and mostpoorly paid blue-collar job category. From 1960 to 2005, the share ofwomen among laborers increased by a factor of ten, going from 2percent to 20 percent. Women have gained access to blue-collarjobs, but only those at the bottom of the occupational distribution.Within the blue-collar sector in general, women remain substan-tially underrepresented compared with their counterparts in thewhite-collar workforce.


236

Figure 6.9 Women in Major Occupational Groups

90

80

70

60

50

40

30

20

10

0

Perc

enta

ge F

emal

e

Profes

sional

Man

ager

ial

Clerica

l

Serv

ices

Sales

Crafts

Operati

ves

Farm

worker

s

Laborer

s

Occupation

19601980

20002005

Source: U.S. Bureau of the Census, Statistical Abstract of the United States.

Stratification at Home and AbroadProcesses of categorical stratification occurring along the lines ofrace, class, and gender have combined to make the United Statesthe most unequal among advanced industrial nations.

A common way of measuring inequality is by forming the ratioof the ninetieth to the tenth percentile of the income distribution.Figure 7.1 uses data from Smeeding (2005) to show this ratio for se-lected nations in the Organization for Economic Cooperation andDevelopment (OECD), essentially the “club” for developed nationsaround the world. These ratios reveal the exceptional nature ofAmerican inequality. The American income ratio of 5.45 is well


250

Figure 7.1 Ratio of 90th to 20th Percentile of IncomeDistribution in Selected OECD Nations

6

5.5

5

4.5

4

3.5

3

2.5

2

Rat

io

Finlan

d

Swed

en

Nether

lands

Germ

any

Belgiu

mFra

nce

Switz

erlan

d

Canad

a

Australi

a

United K

ingdom

United St

ates

Country

2.9 2.96 2.983.18 3.31

3.54 3.623.95

4.334.58

5.45

Source: Smeeding (2005).

figure 7.3. Specifically, this figure shows the percentage of variancein personal income explained by each categorical factor from 1950to 2005.

As can be seen, in 1950 most of the explained variance in per-sonal income was accounted for by gender. The predominant cleav-age in American society was thus between male and female work-ers, with gender explaining 74 percent of the variance comparedwith only 18 percent for class (as measured by education) and 4percent for race. Over the next fifty-five years, however, gendersteadily declined as a significant factor in the American stratifica-tion system, with the share of variance explaining stratificationdropping at a rate of 0.8 percent per year from 1950 to 1980, then ac-celerating to 1.4 percent per year between 1950 and 1980 before leveling off in 2000 and 2005. Over the entire period, the share of in-come stratification explained by gender fell from nearly three-quar-ters in 1950 to less than one-quarter in 2005.

What replaced gender in terms of explanatory power was class:the share of income explained by education went from 18 percent in1950 to 62 percent in 2005. The percentage of variance explained byrace fluctuated somewhat over the period but changed little over-

America Unequal

253

White Black Asian Hispanic Other

Male

Female

College Graduate

One to Three Years of College

High School Graduate

Less Than Twelve Years of Schooling

Figure 7.2 Four-by-Five-by-Two Factorial Design for Analysisof Variance in U.S. Income Inequality, 1950 to 2005

Source: Author’s compilation.

all, bottoming out in 1970 when it reached 3 percent and then risingto 8 percent in 2000 before falling back to 7 percent in 2005. Consis-tent with the conclusions of chapter 6, the interaction between classand gender also increased in importance over the period, as shownin figure 7.4, which plots the percentage of variance captured by thevarious interactions between race, class, and gender from 1950 to2005. Although the percentage of variance explained by the class-gender interaction is small compared to the main effects, its relativeimportance nonetheless rose steadily and significantly from 1950through 2000, going from 3 percent to 7.5 percent before droppingto 5.5 percent in 2005.

Figure 7.5 repeats the ANOVA using family income instead ofpersonal income. Although the overall story is similar, the timing of


254

Figure 7.3 Variance in Personal Income Explained by Race,Education, and Gender, 1950 to 2005

Source: IPUMS.

80

70

60

50

40

30

20

10

0

Perc

enta

ge

1950 1960 1970 1980 1990 2000 2005

Year

GenderRaceEducation

the changes in categorical importance is different when they are fil-tered through the American family system, which itself underwentsignificant change over the period. The critical year in accountingfor family income appears to be 1980. Until that date, the relativeimportance of gender was rising, with the percentage of varianceexplained going from 38 percent in 1950 to 51 percent in 1980, whilethe share explained by education fluctuated around 40 percent withno clear trend. After 1980, however, the relative importance of gen-der in explaining family income dropped markedly, going from 51percent to 28 percent by 2005, while that explained by educationrose from 40 percent to 61 percent.

Figure 7.6 shows the share of variance in family income ex-

America Unequal

255

Figure 7.4 Variance in Personal Income Explained by Race-Education-Gender Interactions

8

7

6

5

4

3

2

1

0

Perc

enta

ge

1950 1960 1970 1980 1990 2000 2005

Year

Gender EducationGender RaceRace EducationGender Race Education

Source: IPUMS.

plained by the various interactions of race, class, and gender at dif-ferent points in time. As with personal income, there is a very clearincrease in the gender-class interaction, but most of the change incategorical importance is observed after 1980, and the decline from2000 to 2005 is sharper. Unlike the case of personal income, how-ever, there is also evidence of an increase in the race-class interac-tion after 1980, though in absolute terms it is relatively small.

The Future of Categorical InequalityAmerican income inequality has thus shifted from stark divisionson the basis of gender to new foundations based on class. That class-based categorical mechanisms have risen to undergird America’s


256

70

60

50

40

30

20

10

0

Perc

enta

ge

1950 1960 1970 1980 1990 2000 2005

Year

GenderRaceEducation

Figure 7.5 Variance in Family Income Explained by Race,Education, and Gender

Source: IPUMS.

new system of inequality is also suggested by figure 7.7, whichshows Gini indices of income inequality for selected OECD nationscomputed before and after the deduction of taxes. The before-taxGini coefficients indicate the amount of raw inequality generated bycapitalist markets in each country, whereas the after-tax Gini indi-cates the degree of inequality that remains after transfers have beenmade to fund the social welfare system in each country. As the graybars show, before taxes the United States is no more or no less equalthan any other country. Its pre-tax Gini is .45, only slightly above thevalue of .44 in Sweden, .43 in Germany, and .42 in the Netherlands,but well below the values of .50 and .49 in Belgium and France andthe same as those prevailing in Australia and Britain.

Taxation works in a redistributive direction for all countries, as

America Unequal

257

Figure 7.6 Variance in Family Income Explained by Race-Education-Gender Interactions

Source: IPUMS.

3.0

2.5

2.0

1.5

1.0

0.5

0.0

Perc

enta

ge

1950 1960 1970 1980 1990 2000 2005

Year

Gender EducationGender RaceRace EducationGender Race Education

indicated by the universally lower values for the after-tax Ginis, butthe extent of the redistribution is much less in the United States. Itscoefficient drops much less as a result of taxation than happens inother countries. The Gini coefficient for income inequality in Bel-gium, for example, falls from .50 before taxes to just .26 after taxes.The after-tax Gini coefficient is at .30 or below for all countries ex-cept Australia, the United Kingdom, and the United States, and inthe former two nations the respective figures are .31 and .34, com-pared with .37 in the United States. Thus, whereas the economy ofBelgium produces a distribution of income that is 11 percent moreunequal than in the United States before taxes, afterward the insti-


258

Figure 7.7 Gini Index of Income Inequality Before and AfterTaxes in Selected OECD Nations

0.5

0.45

0.4

0.35

0.3

0.25

0.2

Gin

i Coe

ffic

ient

Finlan

d

Swed

en

Nether

lands

Germ

any

Belgiu

mFra

nce

Switz

erlan

d

Canad

a

Australi

a

United K

ingdom

United St

ates

Country

0.240.25 0.25 0.25 0.26

0.290.3 0.3

0.31

0.34

0.37

0.450.450.45

0.41

0.39

0.490.5

0.430.42

0.44

0.4

Pretax GiniAfter-tax Gini

Source: Smeeding (2005).

categorically unequal figure 1.1 the stereotype content model

Documents