when work disappears: manufacturing decline and the ... cep ifs.pdf · the decline of us...

When Work Disappears: Manufacturing Declineand the Falling Marriage-Market Value of Men

David Autor 1 David Dorn 2 Gordon Hanson 3

1MIT and NBER

2University of Zurich and CEPR

3UCSD and NBER

October 2017

Autor, Dorn & Hanson When Work Disappears: Manufacturing Decline and the Falling Marriage-Market Value of Men 1 / 59

The Retreat From Marriage: Marriage Rates DivergeBetween College/Non-College in Mid-1980s


In 2009, 40% of U.S. Births Were Out of Wedlock, Morethan Twice as Prevalent as 1980

020

4060

80Pe

rcenta

ge of

Birth

s

Total White Black Hispanic

Percentage of Births to Unmarried Women, by Race

1980 1990 2000 2009

Autor and Wasserman, 2013


Steep Rise in Fraction of Children < 18 in Single-HeadedHouseholds

8% 7%12%

26%

32%

39%

10% 11%

19%18%

29%

42%

56%

70%

80%

31%35% 34%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

White White White Black Black Black Hispanic Hispanic Hispanic

>HS HSGrad <HS >HS HSGrad <HS >HS HSGrad <HS

U.S.Children<18LivingwithMotherOnly:1970&20101970 2010

Autor and Wasserman, 2013


Single & Poor: Poverty Far Higher Among Single-HeadedHHs at Every Education Level (2008 data)

8

MARRIAGE: AMERICA’S GREATEST WEAPON AGAINST CHILD POVERTY

Understanding the Cultural Context of Non-Marital Pregnancy and Childbearing

Clearly, the rise in unwed child-bearing and the decline in marriage play a strong role in promoting child poverty and other social ills. Dealing with these issues will require an understanding of the social con-text of non-marital pregnancy and childbearing. The best source of information on this topic is Promises

I Can Keep: Why Poor Mothers Put Motherhood Before Marriage by Kathryn Edin and Maria Kefalas.27

Edin, professor of public policy at Harvard, is the nation’s most distin-guished researcher on low-income single mothers; her findings overturn much conventional wisdom about

“unintended” pregnancy, out-of wed-lock childbearing, and low-income single parents. In popular perception, out-of-wedlock childbearing occurs

as a result of accidental pregnancies among teenage girls who lack access to or knowledge about birth con-trol. This perception is completely inaccurate.

In reality, unwed births rarely involve teenage girls, are almost never caused by a lack of access to birth control, and generally are not the result of purely accidental pregnancies.

■■ As noted previously, only 8 per-cent of non-marital births occur to girls under 18. Non-marital births and pregnancies are phe-nomena that mainly involve young adult men and women.

■■ Research on lower-income women who have become pregnant out-side of marriage (either as minors or adults) reveals that virtually none of these out-of-wedlock pregnancies occurred because of a lack of knowledge about and access to birth control.28

■■ Out-of-wedlock births are gener-ally not the result of purely acci-dental pregnancies. In fact, most women who become pregnant and give birth out of wedlock strongly desire children. Their pregnancies are partially intended or at least not seriously avoided.29

Most Unwed Mothers Strongly Desire Children

Kathryn Edin explains that chil-dren born out of wedlock are “sel-dom conceived by explicit design, yet are rarely a pure accident either.”30

27. Kathryn Edin and Maria Kefalas, Promises I Can Keep: Why Poor Women Put Motherhood Before Marriage (Berkeley and Los Angeles: University of California Press, 2005).

28. Kathryn Edin, Paula England, Emily Fitzgibbons Shafer, and Joanna Reed, “Forming Fragile Families: Was the Baby Planned, Unplanned, or In Between?” in Kathryn Edin and Paula England, eds., Unmarried Couples with Children (New York: Russell Sage Foundation, 2007), pp. 25–54.

29. Ibid.

30. Ibid., p. 7.

0%

10%

20%

30%

40%

50%

60%

70%

Single Married

High School Dropout*

High School Graduate

SomeCollege

College Graduate

58.8%

38.8%

28.7%

10.6%

24.0%

8.9%4.6%

1.8%

CHART 7

* Virtually none of the heads of families who are high school dropouts are minor teenagers.Source: U.S. Census Bureau, American Community Survey, 2005–2009 data.

POVERTY RATE OF FAMILIES BY EDUCATION AND MARITAL STATUS OF HEAD OF HOUSEHOLD

Both Marriage and Education Are Highly Effective in Reducing Child Poverty in the United States

heritage.orgSR 117

Rector, 2012


When Work Disappears, William Julius Wilson, 1996

“A neighborhood in which peo-ple are poor but employed is dif-ferent from a neighborhood inwhich people are poor and job-less. Many of today’s problemsin the inner-city ghettos—crime,family dissolution, welfare, lowlevels of social organization, andso on—are fundamentally a con-sequence of the disappearance ofwork.”


Hillbilly Elegy: A Memoir of Family and Culture in Crisis,J.D. Vance, 2016

“Wilson’s book spoke to me. Iwanted to write him a letter andtell him that he had describedmy home perfectly. That it res-onated so personally is odd, how-ever, because he wasn’t writ-ing about the hillbilly transplantsfrom Appalachia—he was writingabout black people in the innercities.”


Does declining economic stature of males contribute toretreat from marriage, rise in single headedness?

Exploit well-ID’d trade shocks to manufacturingSizable, sustained e↵ects at local labor market (CZ) level

Measure direct labor market consequencesEmployment status: manuf, non-manuf, unemp, NILFAbsolute and relative earnings levels

Document decline of marriageable menIdlenessSex ratioMortalityOther margins of adjustment

Quantify marriage market interactionsMarital status and cohabitationFertility and birth circumstancesChildren’s household structures & poverty


The decline of US manufacturing

Agenda

1 The decline of US manufacturing

2 Empirical approach

3 Labor market impactsManufacturing employmentEmployment to populationEarnings distribution

4 Male declineIdlenessMissing men and marriageable menMortality among young adults

5 Marriage, fertility, HH structure, povertyMarital status, cohabitationFertilityChildren’s household structures, poverty

6 Conclusions



U.S. Manufacturing Employment Fell by 30% in 1990 -2014

0,00

2,000

4,000

6,000

8,000

10,000

12,000

14,000

16,000

18,000

20,000

22,00019

3919

4219

4519

4819

5119

5419

5819

6119

6419

6719

7019

7319

7719

8019

8319

8619

8919

9219

9619

9920

0220

0520

0820

1120

15

U.S. Manufacturing Employment, 1939 - 2016 (1,000s)

1943,16.6 mil

1979,19.7mil 1999,

17.3mil

2007,13.8mil

2010,11.9mil

2016,12.4mil

1990,17.6mil



China’s Historic Rise as a World Manufacturing Power

0

5

10

15

20

perc

ent

1985 1990 1995 2000 2005 2010Year

China Other emerging economiesUSA Germany

Shares of world manufacturing exports

0

5

10

15

20

perc

ent

1985 1990 1995 2000 2005 2010Year



0

5

10

15

20

perc

ent

1985 1990 1995 2000 2005 2010Year





Literature: Labor Market Impact of ‘China Shock’

Sizable impact of ‘China shock’ on U.S. employment,wages

Bernard, Jensen, Schott ‘06; Autor, Dorn, Hanson ‘13; Ebenstein,Harrison, McMillan, Phillips ‘14; Autor, Dorn, Hanson, Song ‘14;Pierce, Schott ‘16

Explains 30-40% of the decline in manufacturing employment from1990 to 2007 (Caliendo, Dvorkin, Parro ‘15; Acemolgu et al. ‘16)

Impacts concentrated in1 Import-exposed industries: Pierce, Schott ’16; Acemolgu et al. ’162 Workers employed in exposed firms: Autor, Dorn, Hanson, Song ’143 Import-exposed local labor markets: Autor, Dorn, Hanson ’13



Two Thirds of Manufacturing Employees are Men

Emp Status: Men 18 - 39 Emp Status: Women 18 - 39

Non-Manuf55.0%

Non-Manuf57.7%

Non-Manuf62.2%

Non-Manuf63.0%

Non-Manuf64.3%

Manuf23.3%

Manuf21.3%

Manuf17.4%

Manuf14.1%

Manuf9.5%

0%

10%

20%

30%

40%

50%

60%

70%

80%

1970 1980 1990 2000 2014

Non-Manuf37.5%

Non-Manuf49.5%

Non-Manuf58.6%

Non-Manuf59.8%

Non-Manuf63.1%

Manuf8.6%

Manuf10.1%

Manuf8.7%

Manuf6.7%

Manuf3.7%

0%

10%

20%

30%

40%

50%

60%

70%

80%

1970 1980 1990 2000 2014



Manufacturing Jobs Have Higher Wage Levels

$4.96

19.0%

8.0%

11.0%

$5.92

26.0%

17.0%

9.0%

0

5

10

15

20

25

30

Annual Wage and Salary Income

($1,000s)Log Annual Wage and Salary Income Log Annual Hours Log Hourly Wage

Men Women

Estimates based on Census 2000 data are conditional on age, education, race, eth-nicity, and CZone fixed e↵ects



Literature: Gender, Earnings, and Marriage

Becker ‘73: Marriage and specialization

Marital matching negatively correlated on earnings

Increases in male relative to female earnings raise marriage (and v.v. forfalls in relative male earnings)

Evidence

Blau, Kahn, Waldfogel ‘00; Ellwood, Jencks ‘04; Murray ‘12; Shenhav‘16; Shaller ‘16

Black, McKinnish, Sanders ‘03, ‘05: Using coal and steel shocks, finde↵ects on welfare receipt, single-headedness

Bertrand, Kamenica, Pan ‘15: Marriages in which wife earns more thanhusband appear to form less frequently, dissolve more frequently

Kearney, Wilson ‘17: Fracking booms increase incomes and fertility butdo not a↵ect marriage

Page, Hu↵ Stevens, Lindo ‘09; Lindo, Hansen, Schaller ‘16: Parentaljob loss and adverse outcomes for children


Empirical approach

Agenda






6 Conclusions


Empirical approach

Data

Trade exposure

UN Comtrade 1991, 2000, 2014: value of goods imports matched to397 manufacturing industries

County Business Patterns 1980, 1990: employment in 397 industries in722 Commuting Zones (CZs)

Outcome variables

Census 1970, 1980, 1990, 2000, ACS 2013-15: employment, earnings,marital status and household structure in 722 CZs

Vital Statistics 1990, 2000, 2010: births and deaths in 722 CZs

Army Enlistment Records 2001-2011

Focus on population age 18-39, all races/ethnicities combined


Empirical approach

Measuring Local Labor Market Import Exposure

1 Compute change in manufacturing import penetration by industry j(397 industries)

�IPj,⌧ =

�Mch,usj,⌧

Yj,91 +M

j,91 � Xj,91

,

where ⌧ 2 {1991� 2000; 2000� 2014}

2 Compute change in import penetration by Commuting Zone (CZ) ibased on i’s initial industry employment mix across industries j

�IPi,⌧ =

X

j

Lij,90

Li,90

�IPj,⌧


Empirical approach

Instrumental Variables Strategy

Source of endogeneity

US imports from China not only a↵ected by Chinese productivitygrowth and falling trade costs, but also by US demand shocks

Instrumental variables approach

Instrument for US imports from China using other developed countries,imports from China (and lags of all other variables)

�IPoth

j,⌧ =�Mch,oth

j,⌧

Yj,88 +M

j,88 � Xj,88

, �IPoth

i,⌧ =X

j

Lij,80

Li,80

�IPoth

j,⌧

Variation in industry-level trade shocks has well-specified origin

�0s in China productivity and market access, causing Chinese exportsto many countries to surge simultaneously

As per Goldsmith-Pinkham et al. ’17, use fixed weights for industryshares, test for pre-trends


Empirical approach

Geographic Dispersion of Exposure to Chinese ImportCompetition

Autor, Dorn and Hanson 2013


Empirical approach

Pooled (M+F) Trade Shocks: Employment-Weighted �0sin CZ’s Import Penetration per Decade (' % of CZ GDP)

0.54%

0.73%

0.89%

1.01%0.95%

1.15%1.22%

1.30%

0.00%

0.25%

0.50%

0.75%

1.00%

1.25%

1.50%1990-2000 2000-2014

P25 P50 Mean P75


Empirical approach

Accounting for Gender Di↵erences

Gender-specific trade shocks

Trade shocks di↵erentially a↵ect males or females depending onindustries exposed

Instrumental variables approach

Multiply CZ-by-industry exposure measure by initial period male orfemale share of employment in each industry-CZ cell

�IPm

i,⌧ =X

j

mij,90Lij,90Li,90

�IPj,⌧ ,

�IP f

i,⌧ =X

j

(1�mij,90) Lij,90Li,90

�IPj,⌧ ,

where mij,90 is the male employment share in industry j in CZ i in 1990


Empirical approach

Employment-Weighted �0s in CZ’s Import Penetration perDecade by Sex

0.4%0.4%

0.2%0.3%

0.5%

0.6%

0.3%0.4%

0.6%

0.7%

0.4%

0.5%

0.7%

0.8%

0.5%0.5%

0.0%

0.2%

0.4%

0.6%

0.8%

1.0%1990-2000 2000-2014 1990-2000 2000-2014

Men Women

P25 P50 Mean P75


Empirical approach

Main Estimating Equations

Estimated by 2SLS

�Yi,⌧ = ↵

t

+ �1�IPi,⌧ + X0

i,t� + eit

,

�Yi,⌧ = ↵0

t

+ �0

1�IPm

i,⌧ + �0

2�IP f

i,⌧ + X0

i,t�0 + e0

it

Control vector includes...

Population shares in 5 race/ethnicity, 2 education, and 2 nativity groups

Share of employment in manufacturing, ‘routine-intensive’ occupations,‘o↵shorable’ occupations

Female employment share

Census division dummies

Other specification features

Wage quantile estimates apply Chetverikov, Larsen, Palmer ’16 groupedquantile IV estimator

CZ’s weighted by population, SEs clustered on states


Labor market impacts

Agenda






6 Conclusions


Labor market impacts Manufacturing employment

Agenda






6 Conclusions



OLS and 2SLS Estimates 1990 – 2014: ManufacturingEmployment/Pop, Ages 18-39

Impact on Manufacturing Emp/Pop, Ages 18 - 39

-1.09%

-1.69% -1.64%

-1.05%-0.91%

-1.06%

-2.00%

-1.75%

-1.50%

-1.25%

-1.00%

-0.75%

-0.50%

-0.25%

0.00%

0.25%OLS 2SLS

+Division Dummies

+ Manuf Share

+Occ Controls

+Pop Controls



Reduced Form Test for Pre-Trends: 1970-80, 1980-90,1990-00, 2000-14

Impact on Manufacturing Emp/Pop, Ages 18 - 39

1.69%

0.21%

-1.09%

-0.70%

-2.00%

-1.50%

-1.00%

-0.50%

0.00%

0.50%

1.00%

1.50%

2.00%

2.50%

1970-1980 1980-1990 1990-2000 2000-2014


Labor market impacts Employment to population

Agenda






6 Conclusions



Impact of Manufacturing Shocks on EmploymentStatus/Pop by Sex Ages 18-39, 1990-2014

Impact of Gender Trade Shocks on LF Status by Sex, Ages 18 - 39

-1.54%

-0.88%

0.55%0.36%

0.98%

0.53%

-2.25%

-2.00%

-1.75%

-1.50%

-1.25%

-1.00%

-0.75%

-0.50%

-0.25%

0.00%

0.25%

0.50%

0.75%

1.00%

1.25%

1.50%Men 18 - 39 Women 18 - 39

Employment/Pop Unemployment/Pop NILF/Pop



Impact of Manufacturing Shock on M-F Gap inEmployment Status/Pop, 1990-2014

Impact of Pooled Trade Shock on M-F � in LF Status, Ages 18 - 39

-0.65%

0.19%

0.46%

-1.25%

-1.00%

-0.75%

-0.50%

-0.25%

0.00%

0.25%

0.50%

0.75%

1.00%Employment/Pop Unemployment/Pop NILF/Pop



Impact of Sex-Specific Shocks on Emp/Pop by Sex,1990-2014

-1.5%

-0.9%

-3.1%

0.1%0.2%

-2.0%

-4.5%

-4.0%

-3.5%

-3.0%

-2.5%

-2.0%

-1.5%

-1.0%

-0.5%

0.0%

0.5%

1.0%

1.5%Emp/Pop Men Emp/Pop Women

Overall shock Male shock Female shock


Labor market impacts Earnings distribution

Agenda






6 Conclusions



Impact of Pooled Manufacturing Shock on CZ-Level Maleand Female Annual Earnings, 1990 - 2014 (in 2015$)

-6,00

0-5

,000

-4,00

0-3

,000

-2,00

0-1

,000

0Do

llars

(201

5)

0 10 20 30 40 50 60 70 80 90 100Percentile of Income Distribution

Male Earnings Female Earnings

Effect of Overall Trade Shock on Change in Earnings by Gender



Impact of Manufacturing Shock on CZ-Level Male-FemaleAnnual Earnings Gap in US$

−400

0−35

00−3

000−

2500

−200

0−15

00−1

000−5

000

Dolla

rs (2

015)


Effect of Overall Trade Shock on Change in M−F Earnings Gap



Impact of Manufacturing Shock on CZ-Level Male-FemaleAnnual Earnings Gap as a Pct of Male Earnings in 1990

−14

−12

−10

−8−6

−4−2

0Pe

rcen

tage

of M

ale E

arnin

gs


Effect of Overall Trade Shock on Change in M−F Earnings Gap


Male decline

Agenda






6 Conclusions


Male decline Idleness

Agenda






6 Conclusions



What Do Non-Employed Young Adults Do?

Trade-o↵ between employment opportunities andeducational investments

Trade booms and busts: Atkin ’16; Greenland, Lopresti ’16

Housing booms and busts: Charles, Hurst, Notowidigdo ’16; AparicioFenoll ’16

Growing demand for leisure

Young men increasingly devote time to video games: Aguiar, Bils,Charles, Hurst ’16



Focusing on Young Adults 18-25, a Sharp Rise in Male‘Idleness’ in Trade-Exposed CZs

-1.50%

0.72% 0.79%

-0.87%

0.74%

0.13%

-2.5%

-2.0%

-1.5%

-1.0%

-0.5%

0.0%

0.5%

1.0%

1.5%

Employed In School, Not Employed

Not in School, Not Employed

Men Women



Impact of Manufacturing Shock on ‘Idleness’ Gap AmongYoung Adults 18-25: Males - Females

-0.64%

0.66%

-1.50%

-1.25%

-1.00%

-0.75%

-0.50%

-0.25%

0.00%

0.25%

0.50%

0.75%

1.00%

1.25%

Employed In School, Not Employed

Not in School, Not Employed


Male decline Missing men and marriageable men

Agenda






6 Conclusions



Impact of Manufacturing Shock on Ratio ofMales/Females Ages 18 - 25/39, 1990-2014

Impact of Trade Shocks on M/F Gender Ratio in CZ, Ages 18 - 25/39

-1.02

-1.28

-2.75%

-2.50%

-2.25%

-2.00%

-1.75%

-1.50%

-1.25%

-1.00%

-0.75%

-0.50%

-0.25%

0.00%Age 18-39 Age 18-25



Why Does the M/F Ratio Fall in Trade-Exposed CZs?

1 Di↵erential male migrationBartik ’17 finds small negative overall pop inflow response, no outflowresponseNo evidence on gender-specific migration patterns

2 Military enlistmentUsing Army records: sharp rise in enlistments of young, unmarried menCan explain about 10% of M/F decline

3 IncarcerationDeiana ‘15, Feler-Senses ‘15: increase in property crimeDi�cult to translate into population e↵ects w/o incarceration data

4 HomelessnessDi�cult to quantify. Some data suggest that 2/3rds of homeless are men

5 MortalityCase & Deaton ‘15,‘16: Rise in mortality among middle-age,non-Hispanic whites 1998-2015. See also Pierce-Schott ‘17We focus on young adults 20-39, by sex and cause


Male decline Mortality among young adults

Agenda






6 Conclusions



Impact of Manufacturing Shock on ‘Deaths of Despair’

Deaths per 100K Adults Ages 20-39

2.3

-0.1 -0.2 -0.2

0.20.1

-0.2 -0.1

0.00.4

-4/100K

-3/100K

-2/100K

-1/100K

0/100K

1/100K

2/100K

3/100K

4/100K

5/100K

Drug/ Alc Poison

Liver Disease Diabetes

Lung Cancer Suicide

Male Mortality Female Mortality

Annual death rate among Men in 1990 was 213/100K

Annual death rate among Women in 1990 was 79/100K



Male-Female Mortality Di↵erential

5.4

3.02.5

-4/100K

-2/100K

0/100K

2/100K

4/100K

6/100K

8/100K

10/100K

12/100K

All Causes Other Drug + Alcohol

Poisoning

Liver Disease

Diabetes Lung Cancer

Suicide



Summary of Outcomes at Individual (not Family) Level forYoung Adults

Trade shocks...1 Reduce di↵erentially reduce male employment

2 Reduce male relative earnings, esp. below the median of theannual earnings distribution

3 Sharply increase ‘idleness’ among young men but not women

4 Reduce the ratio of young adult men to women in CZ

5 Induce di↵erential rise in male mortality from drug & alcoholcauses

6 Found to raise local crime–mostly committed by young men

Young men faring di↵erentially poorly in trade-impacted CZs


Marriage, fertility, HH structure, poverty

Agenda






6 Conclusions


Marriage, fertility, HH structure, poverty Marital status, cohabitation

Agenda






6 Conclusions



Impact of Manufacturing Shock on Marital Status,Cohabitation of Women Ages 18 – 39, 1990-2014

Impact on Marital Status, HH Structure, Women Ages 18 - 39

-0.95%

-0.21%

1.16%

-0.81%

-0.22%

1.03%

-2.0%

-1.5%

-1.0%

-0.5%

0.0%

0.5%

1.0%

1.5%

2.0%Married

Widowed/ Divorced/ Separated

Never Married

Living w/Spouse

Living w/Partner Other

Marital Status Household Structure



Impact of Manufacturing Shock by Sex on Marital Status,HH Structure of Women Ages 18 – 39, 1990-2014

Marital Status Cohabitation

-3.6%

-0.7%

4.2%

2.0%

0.3%

-2.3%

-5.0%

-4.0%

-3.0%

-2.0%

-1.0%

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%Married

Widowed/ Divorced/ Separated Never Married

Male shock Female shock

-3.2%

0.0%

3.2%

1.9%

-0.5%

-1.4%

-5.0%

-4.0%

-3.0%

-2.0%

-1.0%

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%Living w/Spouse Living w/Partner Other

Male shock Female shock


Marriage, fertility, HH structure, poverty Fertility

Agenda






6 Conclusions


Marriage, fertility, HH structure, poverty Fertility

Impact of Manufacturing Shock on Fertility among WomenAge 15-39, 1990-2014

-2.3

0.4% 0.4%

-7.9

1.2%

2.5%3.5

-0.5%

-1.9%

-12

-10

-8

-6

-4

-2

0

2

4

6

Birth Rate per 1000 Women Age 15-39

Share Teenage Mothers

Share Unmarried Mothers

Overall shock Male shock Female shock


Marriage, fertility, HH structure, poverty Children’s household structures, poverty

Agenda






6 Conclusions



Trade Shocks Increase Prevalence of Children<17 inSingle-Headed and Poor HH’s

Fraction of Children <18 Living in HH Arrangement

0.61%

-0.35%

-0.11%

0.30%0.19%

-0.03%

-0.75%

-0.50%

-0.25%

0.00%

0.25%

0.50%

0.75%

1.00%

1.25%

Poor HH Married Parents

Parent w/ Partner

Single Parent

Grand-parents

Any Other



Shocks to Male vs. Female Earnings Have OppositeImpacts on Children’s HH Arrangements and Poverty

Fraction of Children <18 Living in HH Arrangement

2.1%

-1.9%

0.3%

1.4%

0.4%

-0.3%

-1.1%

1.4%

-0.6%

-1.0%

-0.1%

0.3%

-3.0%

-2.5%

-2.0%

-1.5%

-1.0%

-0.5%

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%Poor HH

Married Parents

Parent w/ Partner

Single Parent

Grand-parents Any Other

Male Shock Female Shock


Conclusions

Agenda






6 Conclusions


Conclusions

The Declining Economic Stature of Young Men and the‘Retreat from Marriage’

1 Decline of manufacturing due to increased import competitionReduced relative emplyoment and earnings of men

2 Broader consequences on young adults and childrenReduced male/female ratio in non-institutional populationRaised male mortality due to drug and alcohol poisoningReduced marriage rates and fertilityRaised fraction of kids living in poverty, single-headed HHs

3 Mechanism appears quantitatively relevantChina shock alone explains 5%-20% of the observed changes in familystructure outcomesTotal e↵ect of manufacturing decline is likely larger


when work disappears: manufacturing decline and the ... cep ifs.pdf · the decline of us...

Documents