social security and trends in inequality

51
0/38 Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion Social Security and Trends in Inequality Sylvain Catherine Max Miller Natasha Sarin Wharton Wharton PennLaw & Wharton

Upload: others

Post on 28-Dec-2021

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Social Security and Trends in Inequality

0/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Social Security and Trends in Inequality

Sylvain Catherine Max Miller Natasha SarinWharton Wharton PennLaw & Wharton

Page 2: Social Security and Trends in Inequality

1/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Motivation – Top 1% wealth share

1920 1930 1940 1950 1960 1970 1980 1990 2000 20100

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Page 3: Social Security and Trends in Inequality

1/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Motivation – Top 1% wealth share

1920 1930 1940 1950 1960 1970 1980 1990 2000 20100

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Is inequality as bad as in 1920?

Page 4: Social Security and Trends in Inequality

1/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Motivation – Top 1% wealth share

1920 1930 1940 1950 1960 1970 1980 1990 2000 20100

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Is inequality as bad as in 1920?

Social Security,

Unemployment InsuranceMedicare,

Medicaid

Expansion of

Social Security

Page 5: Social Security and Trends in Inequality

2/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

This Paper

• Compute aggregate Social Security wealth

– Present value of future benefits, net of future taxes

– Based on Survey of Consumer Finances (SCF) for retirees

– Using Monte Carlo simulations for working households

• Distribute aggregate Social Security wealth between bottom 90% and top 10%

• Recompute the evolution of top wealth shares between 1989-2016

Page 6: Social Security and Trends in Inequality

2/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

This Paper

• Compute aggregate Social Security wealth

– Present value of future benefits, net of future taxes

– Based on Survey of Consumer Finances (SCF) for retirees

– Using Monte Carlo simulations for working households

• Distribute aggregate Social Security wealth between bottom 90% and top 10%

• Recompute the evolution of top wealth shares between 1989-2016

Page 7: Social Security and Trends in Inequality

2/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

This Paper

• Compute aggregate Social Security wealth

– Present value of future benefits, net of future taxes

– Based on Survey of Consumer Finances (SCF) for retirees

– Using Monte Carlo simulations for working households

• Distribute aggregate Social Security wealth between bottom 90% and top 10%

• Recompute the evolution of top wealth shares between 1989-2016

Page 8: Social Security and Trends in Inequality

3/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Key finding – Top 1% wealth share

1920 1930 1940 1950 1960 1970 1980 1990 2000 20100

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Including

Social Security

Page 9: Social Security and Trends in Inequality

4/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

How Does Social Security Work?

Page 10: Social Security and Trends in Inequality

5/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

How Does Social Security Work?

• Paying in

– 12.4% payoll tax: 10.6% to old-age program, 1.8% as disability

– NB: we exclude disability, conservative

– Up to cap (2019 $132,900)

• Benefits replace higher share for lower-wage workers

– Take best 35 years and adjust for inflation and real wage growth

– Replace

– 90% of AIYE below first bend point (2019: $11,112)

– 32% between first and second (2019: $66,996)

– 15% above the second

Page 11: Social Security and Trends in Inequality

5/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

How Does Social Security Work?

• Paying in

– 12.4% payoll tax: 10.6% to old-age program, 1.8% as disability

– NB: we exclude disability, conservative

– Up to cap (2019 $132,900)

• Benefits replace higher share for lower-wage workers

– Take best 35 years and adjust for inflation and real wage growth

– Replace

– 90% of AIYE below first bend point (2019: $11,112)

– 32% between first and second (2019: $66,996)

– 15% above the second

Page 12: Social Security and Trends in Inequality

6/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Stylized Facts:

Why does Social Security matter?

Page 13: Social Security and Trends in Inequality

7/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Social Security promises are worth more than $30trSSA Estimates of NPV of Social Security Promises

01

02

03

04

02

01

8 D

olla

rs,

Trillio

ns

1985 1995 2005 2015

Page 14: Social Security and Trends in Inequality

8/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Social Security benefits are fairly evenly distributed

0.1

.2.3

.4.5

.6In

co

me

Sh

are

1 2 3 4 5 6 7 8 9 10

Decile of Marketable Wealth

Social Security Benefits

Capital Income

Page 15: Social Security and Trends in Inequality

9/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Stylized Facts:

Why did aggregate Social Security wealth increase?

Page 16: Social Security and Trends in Inequality

10/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Social Security wage base increased

03

0,0

00

60

,00

09

0,0

00

12

0,0

00

20

18

Do

llars

1965 1975 1985 1995 2005 2015

Benefits Base

Wage Index

A. Benefits base and Wage index

01

23

x W

ag

e in

de

x

1965 1975 1985 1995 2005 2015

B. Benefits base/Wage index

Earnings exempt fromSocial Security

Page 17: Social Security and Trends in Inequality

11/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Discount rates declined

Market Implied Nominal Yield Curve0

.03

.06

.09

An

nu

aliz

ed

Sp

ot

Ra

te

0 20 40 60 80 100Horizon

2016 2001 1989

Page 18: Social Security and Trends in Inequality

12/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Boomers are reaching retirement age0

.1.2

Sh

are

of

Po

pu

latio

n

0−9 10−19 20−29 30−39 40−49 50−59 60−69 70−79 80−89 90+

1989 2016

Page 19: Social Security and Trends in Inequality

13/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Methodology

Page 20: Social Security and Trends in Inequality

14/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Defining Social Security wealth• Net present value of Social Security

Social Security Wealthit =T∑

s=c+66

(s−1∏k=t

(1−mitk)

)E [Benefitsit](1 + rts)s−t

−c+65∑

s=t+1

(s−1∏k=t

(1−mitk)

)E [Taxesit]

(1 + rts)s−t

– rts: market yield curve in year t– mitk: mortality rates for year t

• For retirees

Social Security Wealthit =T∑

s=t

(s−1∏k=t

(1−mitk)

)Benefitsit

(1 + rt,s)s−t

E[CPIs]CPIt

– Benefits are observed in the data

Page 21: Social Security and Trends in Inequality

14/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Defining Social Security wealth• Net present value of Social Security

Social Security Wealthit =T∑

s=c+66

(s−1∏k=t

(1−mitk)

)E [Benefitsit](1 + rts)s−t

−c+65∑

s=t+1

(s−1∏k=t

(1−mitk)

)E [Taxesit]

(1 + rts)s−t

– rts: market yield curve in year t– mitk: mortality rates for year t

• For retirees

Social Security Wealthit =T∑

s=t

(s−1∏k=t

(1−mitk)

)Benefitsit

(1 + rt,s)s−t

E[CPIs]CPIt

– Benefits are observed in the data

Page 22: Social Security and Trends in Inequality

15/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Social Security wealth of workers

• For each SCF survey year, we simulate mean Social Security wealth for each cohort andgender

• Simulating past and future earnings trajectories:

– Stochastic component: rich process estimated in Guvenen et al. (2019), which matches momentsfrom the cross-section and dynamics of earnings

– Life-cycle component: matches earnings per cohort×gender×year reported in Guvenen et al.(2018)

– Goal: emulating Social Security administrative panel data

• For each simulated path, we discount future benefits net of future taxes

Page 23: Social Security and Trends in Inequality

15/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Social Security wealth of workers

• For each SCF survey year, we simulate mean Social Security wealth for each cohort andgender

• Simulating past and future earnings trajectories:

– Stochastic component: rich process estimated in Guvenen et al. (2019), which matches momentsfrom the cross-section and dynamics of earnings

– Life-cycle component: matches earnings per cohort×gender×year reported in Guvenen et al.(2018)

– Goal: emulating Social Security administrative panel data

• For each simulated path, we discount future benefits net of future taxes

Page 24: Social Security and Trends in Inequality

15/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Social Security wealth of workers

• For each SCF survey year, we simulate mean Social Security wealth for each cohort andgender

• Simulating past and future earnings trajectories:

– Stochastic component: rich process estimated in Guvenen et al. (2019), which matches momentsfrom the cross-section and dynamics of earnings

– Life-cycle component: matches earnings per cohort×gender×year reported in Guvenen et al.(2018)

– Goal: emulating Social Security administrative panel data

• For each simulated path, we discount future benefits net of future taxes

Page 25: Social Security and Trends in Inequality

16/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Calibration & Aggregation

• Social Security parameters

– We assume that parameters of Social Security formula scale up with the wage index– e.g. Earnings cap, bend points

– Consistent with the last 40 years

• Macroeconomic assumptions

– Discount rates: average nominal market yield curves (Fed Board)

– Inflation projections: historical SSA Annual Report

– Real growth rate of wages: historical SSA Annual Report

• Aggregation:

– We merge with the SCF the mean Social Security wealth by age×year×gender group

– We aggregate using SCF survey weight

Page 26: Social Security and Trends in Inequality

17/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Validation – Simulated vs actual full-retirement-age benefits0

10

,00

02

0,0

00

1989 1995 2001 2007 2013

A. Men

Simulated

Data

01

0,0

00

20

,00

01989 1995 2001 2007 2013

B. Women

Page 27: Social Security and Trends in Inequality

18/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Validation – Aggregate Social Security wealth0

10

20

30

40

20

18

Do

lla

rs,

Trillio

ns

1989 1992 1995 1998 2001 2004 2007 2010 2013 2016

Social Security

Actuaries

Page 28: Social Security and Trends in Inequality

18/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Validation – Aggregate Social Security wealth0

10

20

30

40

20

18

Do

lla

rs,

Trillio

ns

1989 1992 1995 1998 2001 2004 2007 2010 2013 2016

Social Security

Actuaries

Our valuation using

SSA discount rates

Page 29: Social Security and Trends in Inequality

18/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Validation – Aggregate Social Security wealth0

10

20

30

40

20

18

Do

lla

rs,

Trillio

ns

1989 1992 1995 1998 2001 2004 2007 2010 2013 2016

Our valuation using real yield curve

(Treasury Inflation Protected Securities)

Page 30: Social Security and Trends in Inequality

18/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Validation – Aggregate Social Security wealth0

10

20

30

40

20

18

Do

lla

rs,

Trillio

ns

1989 1992 1995 1998 2001 2004 2007 2010 2013 2016

Our valuation using real yield curve

(Treasury Inflation Protected Securities)

Our valuation using the

nominal yield curve

and SSA inflation forecasts

Page 31: Social Security and Trends in Inequality

19/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Assigning Social Security wealth

Social Security wealth by age in 20160

.51

1.5

Dol

lars

, Tril

lions

20 30 40 50 60 70 80 90Age

1. Simulation: aggregate Social Security wealth of 45 year-olds in 2016

• $555 billion

Page 32: Social Security and Trends in Inequality

20/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Assigning Social Security wealth

Share of cohort in the Top 10%0

1020

30Pe

rcen

t

20 30 40 50 60 70 80 90Age

Top 1%Rest of Top 10%

2. SCF: To be in Top 10% overall, a 45 year-old need to be in the

• Top 5% of his cohort

Page 33: Social Security and Trends in Inequality

21/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Assigning Social Security wealth

Wealthiest retirees’ share of Social Security wealth φt(x)0

.1.2

.3.4

.5.6

.7.8

.91

Shar

e of

Soc

ial S

ecur

ity W

ealth

0 20 40 60 80 100Wealthiest x%

3. SCF: for young retirees, share of Social Security wealth of top 5%

• 6.5%

Page 34: Social Security and Trends in Inequality

22/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Assigning Social Security wealth

Social Security wealth by age in 20160

.51

1.5

Dol

lars

, Tril

lions

20 30 40 50 60 70 80 90Age

Top 1%Rest of Top 10%Bottom 90%

4. Split of Social Security wealth at age 45 in 2016

• 6.5% x $555 billion = $36 billion for top 10%• 93.5% x $555 billion = $519 billion for bottom 90%

Page 35: Social Security and Trends in Inequality

23/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Top wealth shares0

.1.2

.3.4

.5.6

.7.8

1989 1995 2001 2007 2013

A. Top 10%

0.1

.2.3

.4.5

.6.7

.8

1989 1995 2001 2007 2013

B. Top 1%

with Social Security without Social Security

Page 36: Social Security and Trends in Inequality

24/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Risk-Adjusted Valuation

Page 37: Social Security and Trends in Inequality

25/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Systematic risk of Social Security (1)

• Returns on PAYG contributions depend on growth rate of population and per-capita earnings(Samuelson (1958))

• For US Social Security, wage-indexation explicitly ties returns to the performance of the labor market

– Tax cap follows wage index

– Parameters of benefit function (bend points) scale up with the wage index

→ Before age 60, +1% to wage-index ⇒ +1% to all Social Security cash flows

• Long-run relationship between the labor and stock markets implies that Social Security participantsare exposed to long-run systematic risk of the market portfolio(Geanakoplos and Zeldes (2010) and Catherine (2019))

Page 38: Social Security and Trends in Inequality

26/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Aggregate risk-adjusted valuation0

10

20

30

40

20

18

Do

lla

rs,

Trillio

ns

1989 1992 1995 1998 2001 2004 2007 2010 2013 2016

Page 39: Social Security and Trends in Inequality

27/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Top wealth shares0

.1.2

.3.4

.5.6

.7.8

1989 1995 2001 2007 2013

A. Top 10%

0.1

.2.3

.4.5

.6.7

.8

1989 1995 2001 2007 2013

B. Top 1%

Risk−free valuation Risk−adjusted valuation

No Social Security

Page 40: Social Security and Trends in Inequality

28/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Risk-adjusted valuation: Wealth composition over time

Page 41: Social Security and Trends in Inequality

29/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Discussion

• Role of interest rates

• Robustness

– Funding gap

– Life expectancy inequality

– Adjusting previous studies

Page 42: Social Security and Trends in Inequality

30/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Role of interest rates

• Falling rates redistribute wealth from holders of short-duration assets to those withlong-term investments (Auclert (2019))

• Long duration bonds have performed exceptionally well over the last few decades(Binsbergen (2020))

• Social Security is a leveraged position

Replicating portfolio:– Sell short and medium maturity bonds– Buy very long maturity bonds

Share of Social Securitywealth in 1989 Change since 1989

Benefits Taxes Benefits Taxes NPV(a) (b) (c) (d) (a)·(c)+(b)·(d)

Bottom 99% 229% -129% +407% +118% +780%Top 1% 123% -23% +303% +282% +308%

Entire population 225% -125% +405% +119% +763%

Page 43: Social Security and Trends in Inequality

30/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Role of interest rates

• Falling rates redistribute wealth from holders of short-duration assets to those withlong-term investments (Auclert (2019))

• Long duration bonds have performed exceptionally well over the last few decades(Binsbergen (2020))

• Social Security is a leveraged position

Replicating portfolio:– Sell short and medium maturity bonds– Buy very long maturity bonds

Share of Social Securitywealth in 1989 Change since 1989

Benefits Taxes Benefits Taxes NPV(a) (b) (c) (d) (a)·(c)+(b)·(d)

Bottom 99% 229% -129% +407% +118% +780%Top 1% 123% -23% +303% +282% +308%

Entire population 225% -125% +405% +119% +763%

Page 44: Social Security and Trends in Inequality

30/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Role of interest rates

• Falling rates redistribute wealth from holders of short-duration assets to those withlong-term investments (Auclert (2019))

• Long duration bonds have performed exceptionally well over the last few decades(Binsbergen (2020))

• Social Security is a leveraged position

Replicating portfolio:– Sell short and medium maturity bonds– Buy very long maturity bonds

Share of Social Securitywealth in 1989 Change since 1989

Benefits Taxes Benefits Taxes NPV(a) (b) (c) (d) (a)·(c)+(b)·(d)

Bottom 99% 229% -129% +407% +118% +780%Top 1% 123% -23% +303% +282% +308%

Entire population 225% -125% +405% +119% +763%

Page 45: Social Security and Trends in Inequality

31/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Funding gap

Projected funding gap in 2016 SSA Annual Report0

.2.4

.6.8

1S

ha

re o

f S

ch

ed

ule

d B

en

efits

2020 2040 2060 2080 2100 2120

Low cost

Intermediate cost

High cost

Page 46: Social Security and Trends in Inequality

32/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Funding gap: Top shares (risk-adjusted)0

.1.2

.3.4

.5.6

.7.8

1989 1995 2001 2007 2013

A. Top 10%

0.1

.2.3

.4.5

.6.7

.8

1989 1995 2001 2007 2013

B. Top 1%

Low cost Intermediate cost

High cost No Social Security

Page 47: Social Security and Trends in Inequality

33/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Adjusting for differences in life expectancy

01

00

00

02

00

00

03

00

00

0S

ocia

l S

ecu

rity

We

alth

1989 1995 2001 2007 2013

A. Top 10%

No Life Expectancy Adjustment

Life Expectancy Adjustment

01

00

00

02

00

00

03

00

00

0S

ocia

l S

ecu

rity

We

alth

1989 1995 2001 2007 2013

B. Bottom 90%

Page 48: Social Security and Trends in Inequality

34/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Adjusting for differences in life expectancy0

.1.2

.3.4

.5.6

.7.8

1989 1995 2001 2007 2013

A. Top 10%

0.1

.2.3

.4.5

.6.7

.8

1989 1995 2001 2007 2013

B. Top 1%

No Life Expectancy Adjustment Life Expectancy Adjustment

Page 49: Social Security and Trends in Inequality

35/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Adjusting other studies0

.1.2

.3.4

1989 1995 2001 2007 2013

A. Without Social Security

0.1

.2.3

.4

1989 1995 2001 2007 2013

B. With Social Security (Risk−adjusted)

SCF Saez & Zucman (2016)

Smith, Zidar & Zwick (2020) Batty et al. (2019)

Page 50: Social Security and Trends in Inequality

36/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Adding Social Security to Smith, Zidar, and Zwick (2020)0

.1.2

.3.4

1989 1995 2001 2007 2013

A. Top 1%

0.0

5.1

.15

.2

1989 1995 2001 2007 2013

B. Top 0.1%

0.0

25

.05

.075

.1

1989 1995 2001 2007 2013

C. Top 0.01%

Page 51: Social Security and Trends in Inequality

37/38

Introduction How Does Social Security Work? Stylized Facts Methodology Risk-Adjusted valuation Discussion Conclusion

Conclusion

• Saez and Zucman (2016) argue that Social Security should not be taken into account because it wouldcall for the inclusion of other programs that reduce private savings and it would “not be clear where tostop”

• We argue that narrowly defined marketable wealth is not the right place to stop

– Social Security is 57% of the wealth of the bottom 90%

– Social programs can make marketable wealth inequality look worse

– Current wealth inequality measures cannot be used for policy evaluation

• Top wealth shares have not increased since 1989 when Social Security wealth is taken into account