Skills, Education, and the Rise of Earnings Inequality Among the

“Other 99 Percent”

David AutorMIT Department of Economics and NBER

SNS Seminar, Stockholm 28 August 2015

The Other 99%

1. Is it all about the ‘One Percent’?

2. Do skill investments still pay off?

3. Why are skill returns so high?

4. Job ‘polarization’ and the middle class

5. Are we running out of jobs?

6. Does inequality impede mobility?

7. Conclusions

0%

5%

10%

15%

20%

25%19

13

1918

1923

1928

1933

1938

1943

1948

1953

1958

1963

1968

1973

1978

1983

1988

1993

1998

2003

Shar

e of

tota

l inc

ome

accr

uing

to e

ach

grou

p

Figure 2. Decomposing the Top Decile US Income Share into three Groups, 1913–2007Atkinson, Piketty, Saez 2011

Concentration of U.S. Incomes has Risen Dramatically Since the Late 1970s

Top 1% (incomes above $398,900 in 2007)

Top 5–1% (incomes between $155,400 and $398,900)

Top 10–5% (incomes between $109,600 and $155,400)

Comparison of U.S., Canada, Spain, Sweden and Finland

Figure 7. Effect of Capital Gains on Share of Top Percentile, 1949–2006Atkinson, Piketty, Saez 2011

0

5

10

15

20

2519

49

1953

1957

1961

1965

1969

1973

1977

1981

1985

1989

1993

1997

2001

2005

Shar

e in

tota

l inc

ome

ofto

p pe

rcen

tile

(in p

erce

nt)

U.S.

Canada

Spain

Sweden

Finland

U.S. with CGs

Canada with CGs

Spain with CGs

Sweden with CGs

Finland with CGs

Median Earnings Gap b/w College & High School Roughly Doubles between 1979and 2012

College/high school median annual earnings gap, 1979–2012In constant 2012 dollars

0

10,000

30,000

20,000

40,000

50,000

60,000

70,000 dollars

1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012

Household gap$30,298 to $58,249

Male gap$17,411 to $34,969

Female gap$12,887 to $23,280

Autor, 2014

Changes in the 90/10Ratio of Full-Time Male Earnings Across Twelve OECD Countries, 1980-201

3.3 2.4 2.6 2.0 4.1 2.4 2.1 2.2 2.7 2.1 2.7 3.6

Numbers at the base of each bar correspondto the 90/10 earnings ratio in each country in 1980.

−0.3

−0.6

0.0

0.9

0.6

0.3

1.2

1.5

Cha

nge

in R

atio

of 9

0th

Per

cent

ile M

ale

Ear

ning

s to

10t

h P

erce

ntile

Mal

e E

arni

ngs,

198

0−20

11

Fran

ce

Finl

and

Japa

n

Sw

eden

Kor

ea

Ger

man

y

Den

mar

k

Net

herla

nds

Aus

tralia

New

Zeal

and

Uni

ted

Kin

gdom

Uni

ted

Sta

tes

OECD Statistical Abstracts, Machin and Van Reenen 2010

The Other 99%

1. Is it all about the ‘One Percent’?

2. Do skill investments still pay off?

3. Why are skill returns so high?

4. Job ‘polarization’ and the middle class

5. Are we running out of jobs?

6. Does inequality impede mobility?

7. Conclusions

U.S. Has Very High Returns to Skill Relative to Other Industrialized Countries

C ro s s -n a t io n a l d if fe re n c e s in wa g e re t u r n st o s k il ls , 2 0 11– 2 0 13P e rc e n t a g e in c re a s e fo r a o n e s t a n d a rd d e v ia t io ni n c re a s e i n s k i l l

0 5 1 0 15 2 0 2 5 3 0 p e rc e n t

S w e d e nC z e c h R .

N o r w a yI t a l y

D e n m a r kC y p r u s

F i n l a n dB e l g i u m

F r a n c eE s t o n i a

S l o v a k R .A u s t r i aN e t h e r l a n d s

J a p a nP o l a n d

C a n a d aK o r e a U . K .

S p a i n G e r m a n y

I r e l a n dU . S .

E a r n i n g sg a i n

9 5 % c o n f id e n c ein t e r v a l

Autor, 2014

Inclusive of Tuition, College in the U.S. is a Better Investment Now Than it Was 50 Years Ago

Avery and Turner, 2012

213K

368K

261K

385K

439K

582K 590K

129K

198K

138K

225K

284K

387K 370K

Present discounted value of college relative to high school degreenet of tuition, 1965–2008College/high school difference, 2009 dollars

100,000

0

200,000

300,000

400,000

500,000

600,000 dollars

1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

Men

Women

The Other 99%

1. Is it all about the ‘One Percent’?

2. Do skill investments still pay off?

3. Why are skill returns so high?

4. Job ‘polarization’ and the middle class

5. Are we running out of jobs?

6. Does inequality impede mobility?

7. Conclusions

Largest Contributor to Widening Earnings Inequality: Deceleration in Supply of U.S. College Graduates

1520

2530

3540

4550

5560

65Co

llege

shar

eofh

ours

worke

d(%

)

1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012

Males: 0-9 Yrs Experience Females: 0-9 Yrs Experience

Autor, 2014

1982 2005

Males

Females

Largest Contributor to Widening Earnings Inequality: Deceleration in Supply of U.S. College Graduates

Autor, 2014

15

20

25

30

35

40

45

50 percent

1964 1970 1976 1982 1988 1994 2000 2006 2012

35

45

55

65

75

85

95

105 percent

1964 1970 1976 1982 1988 1994 2000 2006 2012

The supplyof collegegraduates andthe U.S.college/high school premium,1963–2012Collegeshareofhoursworked(%),1963–2012:All working-ageadults

Collegeversus highschoolwagegap(%)

BA

PredictedbySupply-DemandModel

MeasuredGap

The Other 99%

1. Is it all about the ‘One Percent’?

2. Do skill investments still pay off?

3. Why are skill returns so high?

4. Job ‘polarization’ and the middle class

5. Are we running out of jobs?

6. Does inequality impede mobility?

7. Conclusions

Tacit versus Explicit Knowledge

Michael Polanyi, The Tacit Dimension, 1966“We can know more than we can tell... The skill of a driver cannot be replacedby a thorough schooling in the theoryof the motorcar; the knowledge I have of my own body differs altogether from the knowledge of its physiology.”

Two implications of Polanyi’s Paradox – One technical, one economic1. (Technical)Cannot automate what we don’t explicitly

understand2. (Economic)Tasks that are not substituted by machinesare

typically complemented by machines

EmploymentPolarization inthe U.S: Changes in Employmentby MajorOccupationalCategory,1979–2012

Auto

r,2015

40

30

20

10

0

−10

−20

100

×lo

gC

han

gein

Em

ploy

men

t1979–1989 1989–1999 1999–2007 2007–2012

−.1

0

.1

100

×C

hang

e in

Em

ploy

men

t Sh

are

0 20 40 60 80 100

SmoothedChanges in U.S. Employment by OccupationalSkil Percentile,1979–2012

Auto

r,2015

1979–1989 1989–1999 1999–2007 2007–2012

EmploymentPolarization inSixteen EuropeanUnionCountries,1993 - 2010

Go

os,M

anning

and

Salo

mo

ns,2014

-14.9%

-12.1% -12.0%-10.9% -10.8% -10.7% -10.6% -10.6% -10.4% -10.3% -9.6%

-8.6% -8.5%-7.6%

-6.7%

-4.9%

-18%

-15%

-12%

-9%

-6%

15%

12%

9%

6%

3%

0%

-3%

18%Low Paying MiddlePaying HighPaying

Factors that Shape Impact ofAutomation on Workers’ Earnings

1. Substitution versus complementarity• Workersbenefit if their job tasks are complemented, but not

if tasks are substituted. Greattime to be a CADD operator, bad time to be a paper and pencildraftsperson.

2. Elasticity of final demand• Can either dampen or amplify the gains from automation.

Thinkof agriculture versus health care.

3. Labor supply elasticity• Can mitigate wage gains. Why wages don’t rise rapidly in food

service and hair cutting…

4. Relevance:• Abstracttask-intensiveversus Manual task-intensive jobs

20

15

10

5

0

−5

100

×Lo

gC

hang

ein

Mea

nFT

FYW

age

0 20 40 60 80Skillpercentile (rankedbyoccupation’s1979meanlogwage)

100

SmoothedChanges in U.S. MeanWages by OccupationalSkil Percentile,1979–2012

Auto

r,2015

1979–1989 1989–1999 1999–2007 2007–2012

ATM Deployment and Bank Teller Employment, 1970– 2010

Dispensing jobsAs more ATMs were installed in the United States, the numberof tellers employed did not drop.(thousands) 700

Sources: Ruggles and others, Integrated Public Use Microdata Series: Version 5.0; Bureau ofLabor Statistics, Occupational Employment Survey; and Bank for International Settlements,Committee on Payment and Settlement Systems, various publications.

Bessen, 2015

200

100

01970

80 90 200010

Tellers employedATMs installed

600

500

400

300

The Other 99%

1. Is it all about the ‘One Percent’?

2. Do skill investments still pay off?

3. Why are skill returns so high?

4. Job ‘polarization’ and the middle class

5. Are we running out of jobs?

6. Does inequality impede mobility?

7. Conclusions

The ‘ProductivityProblem’ of 1964

1964: PresidentJohnson establishes“Blue-RibbonNational Commission on Technology, Automation, and Economic Progress”

The Milwaukee-Matic industrial machining tool, 1963

The Future of Polanyi’s Paradox

• Computerization has progressed…• Driving vehicles• Parsing legal documents• Performing agricultural field labor.

• Polanyi’s paradox remains relevant central…• Explain what has not yet been accomplished

• Illuminates the contours of the technological frontier

• Overcoming Polanyi’s Paradox1. Environmental Control:

• Bows to Polanyi’s paradox2. Machine Learning

• Attempts an end-run around it

EnvironmentalControl --Kiva SystemsOrderFulfilment: RoboticDrive Units Move Shelves to Workers for Picking

Kiva-Worker Interaction:GuidedbyLaser Pointer,Worker Picks Items from Shelvesfor Shipping

Machine Learning

• Machine learning• Inductive reasoning versus procedural programming

• Problem• Cannot program a machine to “simulate” a non-routine

task by following a scripted procedure

• Solution• Program a machine to master the task autonomously by

studying successful examples of the task being carried out by others

Imagesof CatsSuccessfulyRecognizedbyGoogle X Labs Team, Using a NeuralNetworkof 16,000 Processors

Le, Ranzato, Moga, Devin, Chen, Corrado, Dean and Ng, 2012

The Other 99%

1. Is it all about the ‘One Percent’?

2. Do skill investments still pay off?

3. Why are skill returns so high?

4. Job ‘polarization’ and the middle class

5. Are we running out of jobs?

6. Does inequality impede mobility?

7. Conclusions

1.0

1.2

1.4

1.6

1.8

2.0

1964 1968 1972 1976 1980 1984 1988 1992 1996 200020042008 2012

Some college

>Bachelor's degree

Bachelor's degree

High schoolgraduate

High schooldropout

)BA

Changes in real wage levels of full-time U.S. workers by sex and education, 1963–2012Real weekly earnings relative to 1963 (men) Real weekly earnings relative to1963 (women

1964 1968 1972 1976 1980 1984 1988 1992 1996 200020042008 2012

1.0

1.2

1.4

1.6

1.8

2.0

College Grads are Faring Relatively Wellb/c Non-College Workers are Faring Absolutely Worse

Autor, 2014

MEN WOMEN

Countries with High Inequality and High Skill Returns Have Low Intergenerational Mobility

Source: Miles Corak

CanadaAustralia

NewZealand

GermanyGermany

Japan

KingdomItaly

FranceUnitedKingdom

Spain France

United

0.1

0.2

0.3

0.4

0.5

0.1

0.2

0.3

0.4

0.5

20 25 30

Incomeinequality(moreinequality )35

NorwayDenmark

Finland

AustraliaCanada

NewZealand

Italy

100 120 140 160

Collegeearningspremium(men25to34)180

Earnings inequalityandeconomic mobility:cross-national relationshipsGenerationalearningselasticity(highervalues implylowermobility)

Generationalearningselasticity(highervalues implylowermobility)

BA

Sweden

FinlandNorwayDenmark

UnitedStates UnitedStates

Sweden

No Decline in Intergenerational Income Mobility for U.S. Children Born 1971– 1974vs. 1979– 1982

Mean percentile income rank of children at ages 29-30 vs. percentilerank of their parents for three groups of birth cohorts: 1971-1974, 1975-

1978, and 1979-1982Chetty, Hendren, Kline, Saez and Turner 2014

3040

5060

70M

ean

Chi

ld In

com

e R

ank

0 20 40 60Parent Income Rank

80 100

1971−1974 1975−1978 1979−1982

No Decline in Intergenerational ‘Educational Mobility’ for U.S. Children Born 1984– 1987vs. 1991– 1993

Mean percentile income rank of children at ages 29-30 vs. percentile rank of their parents for three groups of birth cohorts:

1984-1987, 1988-1990, and 1991-1993Chetty, Hendren, Kline, Saez and Turner 2014

20%

40%

60%

80%

100%

Per

cent

in C

olle

ge a

t Age

19

0 20 40 60Parent Income Rank

80 100

1984−1987 1988−1990 1991−1993

Changes in U.S. Male Emp/Pop and Real Wage Levels 1979– 2008(Ages 25-39)

WSMCBSMC

WCLGWGTC

OCLG

BHSD

BHSG

BCLG

BGTC

OHSD

O

HSG

OSMC

OGTC

10

5

0

−5

−10

−15

−20

−25

−30

Perc

enta

gePo

intC

hang

ein

Mal

eEm

ploy

men

tto

Popu

latio

nR

ates

Fitted Values

WHSD

W

HSG

−25 −20 −15 −10 −5 0 5 10 15 20

Percentage Change in Male HourlyWages

Other Non−White Black

25 30 35 40

White

Autor and Wasserman 2013

Changes in Female Marriage Rates vs. Changes in Male Earnings 1979 – 2008,Ages 25 – 39

BHSG

WHSD

WHSG

WSMC

BSMC

WCLG

WGTC

BHSD

OSMCBCLG

BGTC

OHSD

OHSG

OCLG

OGTC

−25

−30

−15

−20

−10

0

−5

10

5

Perc

enta

gePo

intC

hang

ein

Fe

mal

eM

arria

geR

ates

−25 −20 −15 −10 −5 0 5 10 15 20

Percentage Change in Male Hourly Wages

Other Non−White Black

25 30 35 40

Fitted ValuesWhite

Autor and Wasserman 2013

Per Capita Enrichment Expenditures on Children ($2008) Top versus Bottom Quartileof Households

1994-5 2005-6

Source: Duncan and Murnane, 2011

3,536

5,650

6,975

8,872

8351,264 1,173 1,315

$0

$2,500

$5,000

$7,500

$10,000

1972-3 1983-4

Top income quintile

Bottomincome quintile

Gapin yearsof Completed Schooling: Studentswithfamily Incomein the Top vs. Bottomquintiles(by Year of Birth-14)

0

Source: Duncan and Murnane, 2011

1

2

3

419

68

1972

1976

1980

1984

1988

1992

1996

Gap

inye

ars

ofco

mpl

eted

scho

olin

g

Fractionof Twelfth-GradersExpectingto Obtaina B.A. by Sex andParents’Education,1979-2007

Source: Brian L. Jacob and Tamara Linkow Wilder, Using data from the Monitoring the Future survey.

Source: Jacob, Linkow and Wilder, 2011

95

90

85

80

75

70

65

60

55

50

45

401976 1981 1986 1991

Males, parents with less than B.A.

Females, parents with less than B.A.

1996 2001 2006

Males, parents with B.A. or higher

Females, parents with B.A. or higher

Perc

entE

xpec

ting

toEa

rna

B.A.

Fractionof Twelfth-GradersExpectingto Obtaina B.A. by Sex andParents’Education,1979-2007

Source: OECD

0.0

0.5

1.0

1.5

Japa

n Tu

rkey

C

hile

Lu

xem

bour

gM

exic

o K

orea

Sw

itzer

land

Uni

ted

Ger

man

y Be

lgiu

m

Fran

ce

Net

herl

ands

U

nite

dSt

ates

Aus

tria

C

zech

New

Zeal

and

Irel

and

Aus

tral

ia

Can

ada

Nor

way

G

reec

e Sl

ovak

Po

rtug

alSp

ain

Isra

el

Pola

nd

Swed

en

Hun

gary

Fi

nlan

dIt

aly

Icel

and

Den

mar

k Es

toni

a Sl

oven

ia

Female/Male Ratio in College Graduation Rates, OECDCountries 2011

2.0

.1 .2 .3 .4CENSUS 2000: Share Families w/ Kids Single Mom

.5

Fraction of Kids w/Single Mom and Income ImmobilityCommuting Zone Level Correlation

Source: Chetty, Hendren, Kline, and Hendren, 2013

Commuting Zones with Higher Fraction of Mother-Headed HH’s Have Lower Mobility

slope

010

2030

4050

The Other 99%

1. Is it all about the ‘One Percent’?

2. Do skill investments still pay off?

3. Why are skill returns so high?

4. Job ‘polarization’ and the middle class

5. Are we running out of jobs?

6. Does inequality impede mobility?

7. Conclusions

The Other 99%1. Encouraging news

• Rising inequality is mostly (not entirely) about skills – Still a meritocracy rather than a pure plutocracy

2. Rising skill returns largely due to two forces• Slowing supply of new college grads after 1980• Secularly rising demand for human expertise, creativity,

adaptability

3. Recent technological leads to job polarization• But does not necessarily raise wages for ‘service’ workers

4. Intergenerational mobility has not (yet) declined• But reason for worry: Labor market→Marriage market→

Adverse impact on children

Types of Jobs that Will Surely Exist in 50 Years

I. Technicaland creative experts and leadersII. Medical professionalsIII. Developers and testers of new ideasIV. Teachers, especially K-12V. Entertainers: Athletes, Musicians, Actors, Chefs,

Comedians

VI. Building and installation workersVII. Skilled repair workersVIII.Personal helpers, coaches,assistants and consultantsIX. Sex workers