The Financial Feasibility of
Delaying Social Security
Gopi Shah Goda
Shanthi Ramnath
John B. Shoven
Sita Nataraj Slavov
SIEPR/Sloan Working Longer Conference
October 8-9, 2015
When to claim Social Security?
Claiming
Age
62
63
64
65
66
67
68
69
70
Own Benefit
(% of PIA)
75.0%
80.0%
86.7%
93.3%
100.0%
108.0%
116.0%
124.0%
132.0%
Own Benefit:
Percent Increase
from 1-Year Delay
6.7%
8.3%
7.7%
7.1%
8.0%
7.4%
6.9%
6.5%
• Delay equivalent to buying annuity.
Spousal Benefit
(% of Primary
PIA)
35.0%
37.5%
41.7%
45.8%
50.0%
50.0%
50.0%
50.0%
50.0%
Spousal Benefit:
Percent Increase
from 1-Year Delay
7.1%
11.1%
10.0%
9.1%
0.0%
0.0%
0.0%
0.0%
Result: Deferring Social Security is
Actuarially Advantageous
• Deferring is a …
–
–
–
–
good deal for single men in average health
better deal for single women in average health
so-so deal for the lower earner in a couple
very good deal for the higher earner in a couple
• Gains from delay increased substantially for 1938 birth
cohort onwards:
– Rule changes
– Mortality improvements
– Historically low interest rates
• Possible to separate retirement and claiming decisions
through use of retirement savings
Actual Claiming Decisions
Source: Health and Retirement Study, Shoven and Slavov (2012), includes
only individuals not working at time of claim
Why don’t individuals delay?
• Many possible explanations:
–
–
–
–
Lack of liquidity or ability to borrow
Private information about mortality
Impatience
Fear that SS will be reformed and benefits will be
reduced
– Do not value the marginal increase in Social Security
annuity from deferring (Fitzpatrick 2014)
– Claim at FRA (Behaghel and Blau, AEJ 2012)
• We explore the first two
IRS Data
• 1940 birth cohort
– Observed from 1999-2011 (ages 59-71)
– IRA balances and withdrawals from 5498 and 1099-R
(aggregate by household)
– Social Security claim year from 1099-SSA
– Wages from W-2 and self employment
• Focus on primary earners claiming retired worker
benefits:
– Drop those receiving disability at any point
– Drop married women
– Drop those claiming before age 62
• Sample of ~1.1 million
Methodology
• Examine first year of IRA distribution versus
first year of SS claim
– Liquidity constraints  withdraw from IRA as
soon as possible
• Examine ratio of IRA balances to average
Social Security benefit by gender/claim age
– Lower bound on feasible length of delay
• Examine mortality of early versus late
claimers.
First Year of Social Security Claim
First Year of Distribution from
Traditional IRA
Difference between Social Security Claim
Year and First Year of IRA Distribution
Percent of Households with IRA Fair Market Value
Greater than Two or Four Years of Social Security
Benefits
Mortality Hazards by Social Security
Claiming Age
HRS: Data
• Panel survey representative of 50+ population
• Sample: Primary earners ~ age 62 who are observed
through normal retirement age
• Birth cohorts 1928 - 1947
• Focus on primary earners claiming retired worker
benefits:
–
–
–
–
Drop those receiving disability/SSI
Drop married women
Drop those with less than 10 years of work
Drop if claim before age 62
• Household wealth: IRAs, defined contribution
pensions, and non-retirement financial assets.
Methodology
• Compare claiming behavior among those with
high versus low wealth
• Compare characteristics of early (before
normal retirement age) versus late (after
normal retirement age) claimers
Summary Statistics: Claiming and
Wealth
Variable
Claim before NRA
Wealth
PIA for gender/cohort
Wealth > 2 x SS benefit for gender/cohort
Wealth > 4 x SS benefit for gender/cohort
IRA balance > 2x SS benefit for gender/cohort
IRA balance > 4x SS benefit for gender/cohort
Birth year
Born in 1938 or later
Obs
4362
4362
4362
4362
4362
4362
4362
4362
4362
Mean
0.72
246641.90
11630.91
0.64
0.54
0.35
0.26
1936.68
0.43
Std. Dev.
0.45
1590488.00
3148.26
0.48
0.50
0.48
0.44
4.48
0.50
Min
0
-1000000
6187.64
0
0
0
0
1928
0
Max
1
9.02E+07
29016.12
1
1
1
1
1947
1
Percent Claiming Early by Wealth and
Birth Cohort
Wealth ≤ 2 years of benefits
Wealth > 2 years of benefits
IRA ≤ 2 years of benefits
IRA > 2 years of benefits
Wealth ≤ 4 years of benefits
Wealth > 4 years of benefits
IRA ≤ 4 years of benefits
IRA > 4 years of benefits
Born before 1938
Born in 1938 or later
Full Sample
68.2%
69.1%
67.2%
71.5%
67.2%
70.1%
67.3%
73.0%
80.4%
74.4%
76.9%
75.7%
80.0%
73.6%
75.9%
78.2%
73.3%
71.4%
71.5%
73.2%
72.7%
71.7%
71.0%
75.2%
Summary Statistics: Individual
Characteristics
Variable
Good health
Spouse in good health
Has long term care insurance
Sposue has long term care insurance
Financial planning horizon ≥ 5 years
P(Cut Social Security)
P(Cut own Social Security)
P(Live to 75)
Obs
4359
2871
4282
2802
4213
3683
3535
4213
Mean
0.84
0.83
0.09
0.08
0.35
61.57
43.51
68.18
Std. Dev.
0.36
0.38
0.28
0.28
0.48
24.07
26.85
23.96
Min
0
0
0
0
0
0
0
0
Max
1
1
1
1
1
100
100
100
Relationship Between Early Claiming
and Individual Characteristics
(1)
(2)
Good
Health
Spouse
Good
Health
-0.0333***
(0.0120)
-0.00480
(0.0157)
0.00631
(0.00932)
0.00214
(0.0119)
-0.0112
(0.0162)
0.0142
(0.891)
-0.104
(1.058)
-2.068**
(0.817)
Dependent variable mean
0.843
0.830
0.088
0.085
0.352
61.572
43.507
68.184
Observations
R-squared
*** p<0.01, ** p<0.05, * p<0.1
4,359
2,871
4,282
2,802
4,213
3,683
0.046
3,535
0.017
4,213
0.035
VARIABLES
Claim before NRA
(3)
(4)
(5)
(6)
(7)
(8)
Spouse Has Financial
Has Long Long Term Planning
P(Cut Own
Term Care
Care
Horizon ≥ 5 P(Cut Social
Social
Insurance Insurance
years
Security)
Security) P(Live to 75)
Conclusions
• Individuals claim Social Security before taking
distributions from IRAs
• Significant share of individuals has sufficient
wealth to delay Social Security by 2-4 years
• Early claimers have higher actual and
subjective mortality, and report worse health