removing the disincentives in social security for long careers october 20, 2006 gopi shah goda,...

35
Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University Sita Slavov, Ph.D., Occidental College

Upload: veronica-bradley

Post on 03-Jan-2016

216 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

Removing the Disincentives in Social Security for Long Careers

October 20, 2006

Gopi Shah Goda, Stanford UniversityJohn Shoven, Ph.D., Stanford UniversitySita Slavov, Ph.D., Occidental College

Page 2: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

2

Motivation

• Life expectancy has improved dramatically since the introduction of Social Security– Period life expectancy for 20-year-olds in

1935 was 66 for males, 69 for females– Today, 20-year-old males have life

expectancy of 76, and females have life expectancy of 80

• However, men are retiring much earlier

Page 3: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

3

Labor Force Participation by Age for Men, 1965 and 2003

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

55 57 59 61 63 65 67 69 71 73 75+

Age

Pe

rce

nt

in L

ab

or

Fo

rce

1965

2003

Early Retirees Retiring 4 Years Younger

3 Years Younger

Page 4: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

4

Labor Force Participation of Men by Remaining Life Expectancy -

Average Length of Retirement Up Almost 50% Since 1965!

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

8 10 12 14 16 18 20 22 24

Remaining Life Expectancy

La

bo

r F

orc

e P

art

icip

ati

on

1965

2003

7.0 Years

5.5 Years

Page 5: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

5

What are implicit Social Security tax rates?

• The implicit Social Security tax rate for an individual age x is defined as the additional taxes paid at age x, less the present value of additional Social Security benefits accrued, divided by earnings at age x

x

gxxx Earnings

NRAxDAVPIAesPayrollTaxecSecTaxRatImplicitSo

),(12

Page 6: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

60

200

400

600

800

1000

1200

1400

1600

1800

2000

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

AIME

PIA

Social Security Benefit Calculation

Earnings History

Cap & Wage Index Earnings

Calculate AIME (Average IndexedMonthly Earnings)

Calculate PIA (Primary Insurance Amount)

Choose Highest 35 Years

slope = 0.90

slope = 0.32

slope = 0.15

Pay CPI-Indexed PIA from Retirement Age until Death

Page 7: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

7

Assumptions

• Single male worker who starts working at age 20, and retires at the NRA

• Assume 2005 benefit rules persist• Earnings by age simulated using a time-series of

average wages for U.S. and an age-wage profile constructed from 2001 and 2002 CPS

• OASI tax rate = 10.6% maximum implicit tax• Real wage growth = 1% • Real discount rate = 2%• Work years are “front-ended”

Page 8: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

8

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 9: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

9

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 10: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

10

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 11: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

11

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Male Mortality

Page 12: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

12

Female Mortality

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 13: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

13

Why do we see these patterns?

• Once a person works 35 years, additional years of earnings are no longer replacing zeroes – they are replacing lower earnings years

• Each additional year of work does not count the same because of the way the system handles progressivity

Page 14: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

14

What do implicit tax rates look like for actual earnings histories?

• Benefits and Earnings Public Use File, 2004• 1% sample of SSA beneficiaries in December

2004• Limit attention to beneficiaries receiving retirement

benefits based on their own earnings histories, and who started working 1951 or later

• This leaves 123,552 individuals in the sample– 75,368 men and 48,184 women– Birth years range from 1910 to 1942

• Earnings above the cap are not available• No way to link couples in the data

Page 15: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

15

Page 16: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

16

Page 17: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

17

Page 18: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

18

Three Possible Reforms

1. Use 40 years rather than 35 in the AIME calculation

2. Disentangle career length and progressivity

zeros) includenot (doesAIME modified

*1

*

adjustmentlengthcareer

*1

*

zeros) (includesAIME

*1

t

t

NRAt

t

NRAt

t

NRA earnw

w

yf

T

yPIAearn

w

w

yT

yfearn

w

w

TfPIA

Current PIA Calculation Proposed PIA Calculation

Page 19: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

19

PIA Under Current and Proposed Law

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

AIME

PIA

PIA formula under Current Law(AIME calculation includes zeros)

PIA formula under Proposed Law for Low, Middle, and High Income Earner(AIME calculation does not include zeros)

Position on ray depends on career length.

Page 20: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

20

Three Possible Reforms

1. Use 40 years rather than 35 in the AIME calculation

2. Disentangle career length and progressivity

3. Establish a “paid-up” category of workers who have worked a full career of 40 years

Page 21: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

21

Monthly Primary Insurance Amount Under Current and Proposed Law

Average Income Earner

$0

$200

$400

$600

$800

$1,000

$1,200

$1,400

10 15 20 25 30 35 40 45

Career Length (years)

2006

Doll

ars

Proposed Law PIA

Current Law PIA

Page 22: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

22

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 23: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

23

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 24: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

24

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 25: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

25

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 26: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

26

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 27: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

27

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 28: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

28

15 20 25 30 35 40 45

-0.1

-0.05

0

0.05

0.1

Career Length

Impl

icit

Tax

Rat

e

average 10th %ile90th %ilefull cap

Page 29: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

29

Page 30: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

30

Page 31: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

31

What impact do these reforms have on progressivity?

Current Law Proposed LawLow Income Earner 2.64% 2.55%Average Income Earner 1.34% 1.34%High Income Earner 0.81% 0.86%

IRRs for Workers with 30-Year Career

Page 32: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

32

What impact do these reforms have on women?

• Since women in the sample are more likely to experience shorter careers, the reforms disproportionately affect women

• Benefit levels are 2.44% higher for males, 0.07% higher for females (vs. 1.52% overall)

• We look at policies to alleviate this redistribution…

Page 33: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

33

What impact do these reforms have on women?

Percentage Change in PIA by Gender and Years of Credit Given for Women

No Credit 1-Year Credit 2-Year Credit 3-Year Credit Male Benefits 2.44% 2.44% 2.44% 2.44% Female Benefits 0.07% 3.29% 6.42% 9.45% Total 1.52% 2.77% 4.00% 5.18%

Page 34: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

34

Summary

• Flat payroll taxes and the current benefit formula together imply that workers face increasing disincentives for working long careers

• This may be contributing to what could be suboptimally long retirement periods

Page 35: Removing the Disincentives in Social Security for Long Careers October 20, 2006 Gopi Shah Goda, Stanford University John Shoven, Ph.D., Stanford University

35

Summary

• Policies that flatten the pattern of implicit taxes as individuals age would reduce the disincentives of working longer careers

• These policies can be enacted in ways that are either revenue- or benefit-neutral in aggregate

• These policies would involve some redistribution from individuals who work short careers to individuals who work long careers