1

Trends In Labor Force Participation

June 2013

Daniel SullivanFederal Reserve Bank of Chicago

2

61%

62%

63%

64%

65%

66%

67%

68%

1987 1992 1997 2002 2007 2012

Labor Force Participation Rate, Trend vs. Actual

CPS DataLF TrendBLS Data

Ages 16+

3

Main Points Participation is trending down for two reasons

– Demographics – we’re getting older– Long-running behavioral trends – participation for most narrow

demographic groups has been dropping steadily over time

Nevertheless, 2012 participation is below its long-term trend by 1.2 percentage points– Even accounting for the high unemployment rate it is 0.67 percentage

points below trend

Groups especially far below trend– The young– Those with low education– Older workers are bucking the trend

4

-1.50%

-0.75%

0.00%

0.75%

1.50%

0

25

50

75

100

16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80

Perc

ent

Age

2012 Labor Force Participation Rates (left axis)MenWomen

2012 Labor Force Participation Rates and Change in Population Share, by Age(percent)

Participation By Age and Sex

Change in Population Share (right axis) 1995-20002010-2015

5

30

35

40

45

50

55

60

65

1977 1982 1987 1992 1997 2002 2007 2012

Ages 16-19(percent)

Labor Force Participation, By Age/Gender

MenWomen

6

60

65

70

75

80

85

90

1977 1982 1987 1992 1997 2002 2007 2012

Ages 20-24(percent)

Labor Force Participation, By Age/Gender

MenWomen

7

80

85

90

95

100

1977 1982 1987 1992 1997 2002 2007 2012

Men, 25-54(percent)

Labor Force Participation, By Age/Gender

25-2930-3435-3940-4445-4950-54

8

50

55

60

65

70

75

80

1977 1982 1987 1992 1997 2002 2007 2012

Women, 25-54(percent)

Labor Force Participation, By Age/Gender

25-2930-3435-3940-4445-4950-54

9

30

35

40

45

50

55

60

65

70

75

80

85

1977 1982 1987 1992 1997 2002 2007 2012

Men, 55-59

Women, 55-59

Men, 60-64

Women, 60-64

Ages 55-64(percent)

Labor Force Participation, By Age/Gender

10

0

5

10

15

20

25

30

35

40

1977 1982 1987 1992 1997 2002 2007 2012

Men, 65-69

Women, 65-69

Men, 70+

Women, 70+

Ages 65+(percent)

Labor Force Participation, By Age/Gender

11

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80

2012 Male Participation Rates By EducationMen

Less than HSHSSome CollegeCollegePost College

12

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80

2012 Female Participation Rates By EducationWomen

Less than HSHSSome CollegeCollegePost College

Forecasting Demographic Group Behavior Question: In 2007, how to forecast participation rates of 50-54 year old

women in 2015?

BLS Method: Extrapolate the historical time series for participation of 50-54 year old women using last 13 years (mixing cohorts)

Cohort Method: – Note that women who will be 50-54 in 2015 were born 1961-65– Compare the LFP of the 1961-65 birth cohorts to those of earlier

cohorts at the same age– Assume cohort differences will persist at higher ages

14

0.9

1

1.1

1.2

1.3

1.4

25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53

Log

Odd

sSelect Model Fit LFP Profile ProjectionsWhite Female HS Graduate, 25-54

1947 Cohort1952 Cohort1957 Cohort1962 Cohort1967 Cohort

15

1955 Birth Cohort

1960 Birth Cohort

Select Model Fit LFP Profiles Through 2007White Female HS Graduate, 25-54

0.9

1

1.1

1.2

1.3

1.4

25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53

Log

Odd

s

1947 Cohort1952 Cohort1957 Cohort1962 Cohort1967 Cohort

Above projections based on extensions of Aaronson and Sullivan, Chicago Fed Economic Perspectives, 2001

Somewhat similar results to Aaronson, Fallick, Figura, Pingle, and Wascher, Brookings, 2006

Methodological differences – Estimates at individual level

(Models estimated using CPS data 1987-2007)– Everything conditional on educational levels– Many details

Cohort-Based Projections

17

Prob individual i of sex s born in year b is in LF at age a

Birth year cohort dummies

Age dummies

Race group dummies

Age-specific controls

Estimated by age groups: 16-19, 20-24, 25-54, 55-70, 71-79.

A Basic Logistic Cohort Model

18

Age-Specific Controls

Ages 16-24– Real Minimum Wage– Hourly Wage Ratio of 16-19 year olds to 25-54 year olds

Ages 25-54– Fraction of population married with a Child 5 Years or Younger– Fraction of population married with no Child 5 Years or Younger

Ages 55 and higher– Gender specific life expectancies

19

Prob individual i of sex s and education e born in year b is in LF at age a

5 education categories: College

Extension: Condition on Education

20

To forecast LFP, need educational attainment forecasts

Prob individual i of sex s born in year b has attainment of at least e at age a given attainment of at least e - 1

Extension: Condition on Education

21

Prob individual i of sex s and education e born in year b is in LF at age a

Annual unemployment gap (actual – CBO NAIRU)

Extension: Allow for Business Cycle Effects

22

Let pt = Overall trend LFP at time t

pdt = Trend LFP for demographic group d at time t

fdt = Share of population in group d at time t

Then

And

BehaviorDemographics

A Decomposition

23

(Percentage points per year)

Decomposition of LFP Change

1987-1997 1997-2005 2005-2010 2010-2013Total Change 0.14 -0.02 -0.16 -0.19

Demographic 0.05 -0.06 -0.08 -0.10

Behavioral 0.09 0.04 -0.07 -0.09

24

1987-1997 1997-2005 2005-2010 2010-2013

Total 0.05 -0.06 -0.08 -0.10

Age 16-19 0.01 0.00 0.01 0.03

Age 20-24 -0.02 0.01 0.00 0.00

Age 25-54 0.05 -0.04 -0.04 -0.05

Age 55-70 0.04 -0.06 -0.07 -0.05

Age 71-79 -0.03 0.03 0.02 -0.03

Decomposition of Demographic Contribution(Percentage points per year)

25

30%

35%

40%

45%

50%

55%

60%

1987 1992 1997 2002 2007 201265%

70%

75%

80%

1987 1992 1997 2002 2007 2012

BLS DataCPS DataLF Trend

16-19 year olds

Labor Force Participation Rate, Trend vs. Actual20-24 year olds

BLS DataCPS DataLF Trend

26

88%

89%

90%

91%

92%

93%

94%

1987 1992 1997 2002 2007 201271%

72%

73%

74%

75%

76%

77%

1987 1992 1997 2002 2007 2012

BLS DataCPS DataLF Trend

Male, 25-54 Female, 25-54

Labor Force Participation Rate, Trend vs. Actual

BLS DataCPS DataLF Trend

27

50%

55%

60%

65%

1987 1992 1997 2002 2007 201230%

35%

40%

45%

50%

55%

1987 1992 1997 2002 2007 2012

Labor Force Participation Rate, Trend vs. Actual

BLS DataCPS DataLF Trend

Male, 55-70 Female, 55-70

BLS DataCPS DataLF Trend

28

0%

5%

10%

15%

20%

25%

1987 1992 1997 2002 2007 20120%

5%

10%

15%

20%

1987 1992 1997 2002 2007 2012

Labor Force Participation Rate, Trend vs. ActualMale, >70 Female, >70

CPS DataLF Trend

CPS DataLF Trend

29

Demographically-Adjusted LFP

63%

64%

65%

66%

67%

68%

1995 2000 2005 2010

16+ CPSAdjusted 16+ LFP

30

1987-1997 1997-2005 2005-2010 2010-2013

Total 0.09 0.04 -0.07 -0.09

Men -0.05 -0.04 -0.08 -0.07

Age 16-19 -0.01 -0.03 -0.04 -0.03

Age 20-24 -0.01 -0.02 -0.01 -0.01

Age 25-54 -0.04 -0.05 -0.05 -0.03

Age 55-70 0.01 0.05 0.01 -0.01

Age 71-79 0.00 0.01 0.01 0.01

Decomposition of Behavioral Contribution(Percentage points per year)

31

1987-1997 1997-2005 2005-2010 2010-2013

Total 0.09 0.04 -0.07 -0.09

Women 0.14 0.08 0.01 -0.02

Age 16-19 0.00 -0.03 -0.02 -0.03

Age 20-24 0.00 -0.01 -0.02 -0.01

Age 25-54 0.08 0.01 -0.02 -0.03

Age 55-70 0.05 0.09 0.04 0.03

Age 71-79 0.01 0.01 0.01 0.02

Decomposition of Behavioral Contribution(Percentage points per year)

32

LFP By Education

42%

44%

46%

48%

50%

1987 1992 1997 2002 2007 201260%

62%

64%

66%

68%

70%

72%

1987 1992 1997 2002 2007 2012

66%

68%

70%

72%

74%

76%

78%

1987 1992 1997 2002 2007 2012

Less than HS HS Degree

Some College

33

LFP By Education

76%

78%

80%

82%

84%

86%

1987 1992 1997 2002 2007 2012

Post-College

76%

78%

80%

82%

84%

1987 1992 1997 2002 2007 2012

College Degree

34

-3.50%

-3.00%

-2.50%

-2.00%

-1.50%

-1.00%

-0.50%

0.00%

0.50%

1.00%

1987 1992 1997 2002 2007 2012

LFP Gap By Education(Actual LF – Predicted LF)

Less than HSHSSome CollegeCollegePost College

35

-0.60%

-0.50%

-0.40%

-0.30%

-0.20%

-0.10%

0.00%

0.10%

0.20%

0.30%

1987 1992 1997 2002 2007 2012

Contribution to LFP Gap By Education(LFP Gap * Population Share)

Less than HSHSSome CollegeCollegePost College

36

Possible Interpretation of Low Education Results

Housing boom may have temporarily stopped the slide of real wages for low education workers …– Possible interpretation of Charles, Hurst, and Notowidigdo (2012)

Temporarily holding up LFP– And our trend estimates

After housing collapse, wages and LFP declined

Another story: Downward nominal wage rigidity bites harder for low education workers– Productivity gains take longer to bring realignment– Probably more a story for unemployment

37

LFP Gap By Age(Actual LF – Predicted LF)

-5%

-4%

-3%

-2%

-1%

0%

1%

2%

3%

4%

1987 1992 1997 2002 2007 2012

16-1920-2425-5455-70>70

38

-0.50%

-0.40%

-0.30%

-0.20%

-0.10%

0.00%

0.10%

0.20%

0.30%

0.40%

1987 1992 1997 2002 2007 2012

Contribution to LFP Gap By Age(LFP Gap * Population Share)

16-1920-2425-5455-70>70

39

Possible Interpretation of Age Results

Young workers most affected by down turn– Consistent with past research that entrants face disproportionate

difficulties in poor labor markets

Young workers may also be returning to school– Understandable given low opportunity costs

Older workers may be working more to compensate for negative shock to wealth

40

Group 2012 LFP Gap Contribution to TotalTotal Gap: -1.14%

Female, 25-54, HS Grads -2.14% -0.14%

Male, 25-54, HS dropouts -3.03% -0.10%

Female, 25-54, Some college -1.03% -0.08%Female, 25-54, HS dropouts -2.74% -0.07%

Male, 20-24, HS graduates -4.07% -0.06%

Female, 71-79, HS Grads -2.34% -0.06%

Male, 25-54, HS Grads -0.71% -0.06%

Male, 25-54, Some College -0.81% -0.05%

Female, 20-24, Some College -1.88% -0.04%

Female, 16-19, Some college -8.67% -0.04%

Male, 16-19, HS dropouts -1.59% -0.04%Residual: -0.40%

(Difference of above gaps to total)

Contribution to 2012 LFP Gap, by sex/age/education

41

Caveats on LFP Modeling

Modeling of business cycle could be improved– E.g., some evidence that LFP responds to unemployment with

very long lags

Could incorporate more effects of policy changes – E.g., on SS, taxes, tuition, etc.

More generally, need better economics– Labor supply responds to wages and other general equilibrium factors

42

Unemployment RatePopulation Growth, Trend vs. Actual(percent)

Unemployment Rate, Natural Rate vs. Actual(percent)

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1987 1992 1997 2002 2007 2012 2017

TrendActual

3

4

5

6

7

8

9

10

1987 1992 1997 2002 2007 2012 2017

TrendActual

43

90

92

94

96

98

1987 1992 1997 2002 2007 2012 20170.50

0.55

0.60

0.65

1987 1992 1997 2002 2007 2012 2017

Payroll EmploymentRatio of Payroll to Household Employment(percent)

Payroll Employment-to-Population Ratio(fraction)

TrendActual

TrendActual

44

Payroll EmploymentPayroll Employment Gap(thousands of jobs)

100000

110000

120000

130000

140000

150000

1987 1992 1997 2002 2007 2012 2017

Avg. monthly gains of 195k close the gap by the end of 2016

TrendActual

45

-50

0

50

100

150

200

250

1988 1993 1998 2003 2008 2013 2018

Payroll EmploymentTrend Payroll Employment Growth(jobs/month)

46

Extra Slides -- May eventually be deleted

47

0

20

40

60

80

100

16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80

Perc

ent

Age

MenWomen

2012 Labor Force Participation Rates, by Age (percent)

Participation By Age and Sex

Change in Population Share, by Age (percentage points)

-0.015

-0.01

-0.005

0

0.005

0.01

0.015

16-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75+

Change 1995-2000Change 2010-2015

48

0

20

40

60

80

100

16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80

Perc

ent

Age

MenWomen

2012 Labor Force Participation Rates(percent)

Participation By Age and Sex

49

Change in Population Share, By Age

-0.015

-0.01

-0.005

0

0.005

0.01

0.015

16-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75+

Change 1995-2000Change 2010-2015

(percentage points)

50

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1987 1992 1997 2002 2007 2012 2017

2009

$

Age-Specific Control VariablesTeen and 20-24 modelsReal Minimum Wage (Demeaned)

51

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

1987 1992 1997 2002 2007 2012 2017

Age-Specific Control VariablesTeen and 20-24 modelsHourly Wage Ratio of 16-19 year olds to 25-54 year olds (Demeaned)

52

8%

9%

10%

11%

12%

1987 1992 1997 2002 2007 2012 2017

Age-Specific Control VariablesPrime age modelsMarried with a Child 5 Years or Younger(percent of 25-54 year olds)

53

20%

22%

24%

26%

28%

1987 1992 1997 2002 2007 2012 2017

Age-Specific Control VariablesPrime age modelsMarried with no Child 5 Years or Younger(percent of 25-54 year olds)

54

26

28

30

32

34

1987 1992 1997 2002 2007 2012 2017

Age-Specific Control VariablesOlder age modelsLife Expectancies by Sex(expected years lived past 50)

MenWomen

55

61%

62%

63%

64%

65%

66%

67%

68%

1987 1992 1997 2002 2007 2012

LF Participation Rate, with Business Cycles

CPS DataLF Trend

Ages 16+

56

Demographically-Adj. LFP, w/Business Cycle Effect

63%

64%

65%

66%

67%

68%

1995 2000 2005 2010

16+ CPSAdjusted 16+ LFP

57

LFP Gap By Education, with Business Cycle Effect(Actual LF – Predicted LF)

Less than HSHSSome CollegeCollegePost College

-2.50%

-2.00%

-1.50%

-1.00%

-0.50%

0.00%

0.50%

1.00%

1.50%

1987 1992 1997 2002 2007 2012

58

-0.40%

-0.30%

-0.20%

-0.10%

0.00%

0.10%

0.20%

0.30%

1987 1992 1997 2002 2007 2012

Contribution to LFP Gap By Education(LFP Gap * Population Share)

Less than HSHSSome CollegeCollegePost College

59-4%

-3%

-2%

-1%

0%

1%

2%

1987 1992 1997 2002 2007 2012

LFP Gap By Age, with Business Cycle Effect(Actual LF – Predicted LF)

16-1920-2425-5455-7071-79

60

Contribution to LFP Gap By Age(LFP Gap * Population Share)

16-1920-2425-5455-7071-79

-0.60%

-0.50%

-0.40%

-0.30%

-0.20%

-0.10%

0.00%

0.10%

0.20%

0.30%

0.40%

1987 1992 1997 2002 2007 2012

61

64%

65%

66%

67%

68%

69%

70%

1987 1992 1997 2002 2007 2012

Labor Force Participation Rate, Trend vs. Actual

CPS DataLF Trend

Estimated with data through 2012, Ages 16-79