Ask the Scholar
Document scope · 1 page
Scholar
Ask about this object, its catalog metadata, its source description, or the page inventory.
For page-specific OCR and visual context, open one of the page chats.
Scholar Source Context
Document identity
localId
24493319
label
Social Security Earnings Test
core
doc
dtoType
document
citationUrl
pageCount
1
Source metadata
id
24493319
sourceUrl
contentType
document
title
Social Security Earnings Test
citationUrl
collections
Records of the First Lady's Office (Clinton Administration)
Shirley Sagawa's Files
imageCount
1
hasImages
yes
source
import
hasTranscription
no
Source extras
naId
24493319
levelOfDescription
fileUnit
otherTitles
42-t-7763274-20130124S-053-005-2015
recordType
description
ocrSource
nara-archive
Single page context
seq
1
pageIndex
0
type
document
mediaId
763b276e4e32c541
ocrText
Social See
Pres: elim can r test F improve benef.t for
pror elderly widows
mts decided House was & Means wid into Pres. proposal-
P imprement needed - WH cause working on
when SOTU
ext. solvery t 2055 - need longer
but or Digartism Losis to jet to 2075
Bill won't have of or benefit changes -
Do more equity? Widows unproved will be
small
what also
fre. sec camp test - aflects 65-69
seemle/ Hl rules affects 62-64
not a tax on easays that releases benefits - 1
at 70, you or when you get a
high benef.' to comperate for lower
benefit earle -
most bells clim 65-69, some 62
Rep. bill HRS chin 65 not 62
mc Cain bill
1.
3
archer I how
62
Shd Dear J f 62m65
By impact on proty on older 3 dows sys. increases
age 62-64- if earnys test llim may more
away elderly widows
Social See
Pres: elin can r test r improve benef.t for
poor elderly widows
mts decided House was L Means wld into Pres. proporal-
of imprement needed - WH cause working on
when SOTU
ext. solvery L 2055 - need longer
but or Dipartism Losis to jet to 2075
Bir won't have 9 or benefit charges -
Do more equity? Widows unproved well be
small
what chico
Src. Dec earys dest - aflects 65-69
seender/ Hl rules affects 62-64
not a tax on easays that reduces benefits 1
at 70, * or when you get a
high benef.' to comperate for love
benefit earlier -
most bells clim 65-69, some 62
Rep. bill HRS chin 65 not 62
mc Cain bill
3
arche Show
62
Shd Dear J f 62n65
By impact on proty on older 3 dows 5ys increases
are 62-64- if earys test llim may nine
away elderly widows
will draw benefits at 62 - additional
people who start drawing at 42 will
set i'. a lower benefit frime yrs
when . 1d w/ no other some 1
income, larger progr I name JUL see
provies- by wellows imes
Peter Orsag * John Gruber at MIT worked $
to do research on easys test -
Then research form that if we clim carn
test before 62- -
different benef to 57 widows 80/- +
syn. $ werease
privaty
)
those w/ reduced b/c drew before 65
are $1500/6 below
there w/o
tid not
poverty
above poraty line
Ohe factor explained 1/2 dill
Research curculated w/in WH -
the Jack, appl etc concluled 65
me reopend b/c WrM Dear stall- want to
follow Other at 62
Person fal must strongly - Ken Apfel
Bill coming from W+M comen abnt 62
CENTER ON BUDGET
AND POLICY PRIORITIES
820 First Street, NE, Suite 510
Washington, DC 20002
Telephone: 202/408-1080
Fax: 202/408-0848 or 408-1056
If there are any problems with the transmission of this document,
please call Betty Hitchcock 202/408-1080
To:
Shirley Sagawa
FAX NUMBER:
456-6244
FROM:
Bob Greenstein
REGARDING:
DATE:
June 22, 1999
NUMBER OF PAGES:
11
(including cover sheet)
Comments:
Shirley please find enclosed the one page memo to Hillary, as well as a short
memo I've written and a memo that Peter Orszag and Jon Gruber wrote to several
senior Administration officials on the Social Security earnings test issue. The copy I
have and am sending you of the Orszag/Gruber memo has the names of the addressees
removed. I think Peter sent it to Gene, Larry Summers, Ken Apfel, Jack Lew, and
maybe a couple of other people.
CENTER ON BUDGET
AND POLICY PRIORITIES
Revised June 21, 1999
SHOULD THE SOCIAL SECURITY EARNINGS TEST BE ELIMINATED
FOR THOSE UNDER AGE 65?
President Clinton and many Members of Congress have called for eliminating
the Social Security "earnings test." There actually are two earnings tests related to Social
Security retirement benefits - one for those 65 through 69 (i.e., those at or above the
age at which full benefits are paid), and another for those under age 65. Each of the
two tests has its own earnings threshold and its own rules for how much Social Security
benefits are reduced when earnings exceed the threshold. A question facing the
Administration and Congress is whether to propose eliminating both earnings tests or
only the test for those aged 65 and over, which is the earnings test that has drawn most
of the criticism.
When policymakers call for eliminating "the earnings test," they can mean either
that they want to eliminate the test for those 65 through 69 or that they want to
eliminate both tests. H.R. 5, the principal House Republican bill to eliminate the
earnings test - which has been introduced by Rep. Sam Johnson and has 83 Republican
cosponsors (including several Ways and Means Committee members) and 16
Democratic cosponsors - eliminates the earnings test for those at or above the full
benefit age. It retains the earnings test for individuals below the full benefit age. The
same is true of both the Stenholm-Kolbe bill and S. 279, a bill to repeal the earnings test
that Senator John McCain has introduced with 10 Republican senators and one
Democratic senator. On the other hand, the Archer-Shaw proposal, also said to
eliminate the earnings test, drops it for beneficiaries age 62 and over.
This memo recommends that the Administration and Congress follow the path
of H.R. 5 and eliminate the test for those at or above the normal retirement age, while
retaining the test for those who elect to draw retirement benefits early. This would be
the better course for several reasons.
Eliminating the earnings test for those under age 65 is very likely to
worsen the extent and depth of poverty among elderly widows.
President Clinton has rightly said we should make changes in the Social
Security benefit structure to reduce poverty among this group.
Eliminating the earnings test for early retirees would do the opposite.
There is no evidence that eliminating the earnings test for those below the
full benefit age will result in more employment among this group.
Evidence of positive effects of eliminating the earnings test on
820 First Street, NE, Suite 510, Washington, DC 20002
I
Tel: 202-408-1080 Fax: 202-408-1056 [email protected] http://www.cbpp.org HN0026
employment is limited and applies only to those at or above the full
benefit age.
Two economists with expertise in Social Security issues, Jon Gruber of M.I.T. and
Peter Orszag, now with the University of California at Berkeley and formerly with the
Council of Economic Advisers and the National Economic Council, have recently
conducted an exhaustive review of research on the earnings test. They also have
undertaken new research in this area. Their findings indicate that eliminating the
earnings test below the full benefit age could have significant adverse effects on
poverty among elderly widows and that there is no evidence to suggest it would boost
employment. A memo they wrote last week to Administration officials last week
recommends, as this memo does, that the earnings test be eliminated for those at or
The Effects on Widows
Currently, about 60 percent of workers begin to collect benefits soon after their 62nd
birthday. Some 80 percent start receiving benefits before they turn 65. If Social Security
benefits for workers in their early 60s are no longer reduced based on earnings, even more
workers will likely opt to begin receiving benefits early even if they continue to work. These
workers will see an increase in their monthly income in the short run (while they are in their
60s and still working) but at the cost of much lower Social Security benefits when they are
older.
Workers who begin to draw benefits before age 65, along with their spouses, receive
actuarially reduced benefits for the rest of their lives. Currently, those who begin to draw
benefits at 62 receive a 20 percent benefit reduction; when the age at which full benefits are
paid rises from 65 to 67, those starting to draw benefits at 62 will face a 30 percent benefit
reduction. Because women tend to live longer than men, the negative impact of these
benefit reductions are sharpest among widows. This means that if the earnings test is
eliminated for those who draw benefits early and, as a consequence, more individuals begin
to draw benefits at 62 while still working, a substantial number of women who outlive their
husbands will be faced with reduced benefits for many years while they are very old.
Eliminating the earnings test for individuals below the full benefit age thus would likely
lead to a significant increase in the number of elderly widows who are poor.
New research by Gruber and Orszag heightens this concern. They examined Social
Security benefits among widows and widowers aged 80 and over in 1996. They found that
the average annual benefit was significantly below the poverty line among those whose
benefits were reduced because they or their deceased spouses had begun to draw benefits
before 65. By contrast, among widows and widowers whose benefits had not been reduced
because of early receipt of benefits, the average annual benefit was substantially higher and
was well above the poverty line. Gruber and Orszag's work suggests that reductions in
monthly benefits caused by decisions to begin receiving benefits before the full benefit age
accounted for about half of this difference.
2
above the full benefit age but not for those below it. (A copy of their memo is
enclosed.)
One other set of issues also should be noted - those relating to cost. Elimination
of the earnings test has no cost over the next 75 years but does have a sizable short-term
cost. Eliminating the test for those at or above the full benefit age would cost about $15
billion over the first five years. Eliminating it for those 62 and over would cost about
$30 billion over the first five years. (In addition, eliminating the earnings test for early
retirees might lead to other measures that would increase Social Security costs over the
full 75-year actuarial balance period. Eliminating the earnings test for early retirees
could pave the way for easing the earnings test for survivors, a test currently tied to the
earnings test for early retirees, and for Social Security disability beneficiaries aged 62 to
64. The estimates which show that eliminating the earnings test has no significant cost
over the 75-year period as a whole do not include the costs that would be incurred if
the earnings tests for survivors and some disabled beneficiaries also were liberalized.)
In short, there are significant drawbacks to eliminating the earnings test for early
retirees, while there do not appear to be countervailing benefits. Eliminating the
earnings test for workers under age 65 would be likely to result in a reduction in the
incomes of a significant number of elderly widows and a rise in the number of such
widows living in poverty, contrary to the President's expressed intention to increase
Social Security benefits and reduce poverty for elderly widows. In addition, there is
neither evidence of positive employment effects from eliminating the earnings test for
early retirees nor a political need to follow such a course. As noted, a number of bills -
including the principal House Republican bill to eliminate the earnings test - drop it
only for those at or above the full benefit age.
Sponsors of virtually all bills that end the earnings test for those at or above the
full benefit age but retain it for early retirees describe their bills as "eliminating the
earnings test." A Democratic Ways and Means Committee bill can and should follow
the same course.
3
To:
From:
Jon Gruber
Peter Orszag
Subject: Earnings Test
Date:
May 17, 1999
We understand from news reports that following your meeting with Ways and Means
Democrats today, a bill will be drafted implementing the President's Social Security reform
proposals. Since that proposal includes elimination of the earnings test in some form, we wanted
to highlight the conclusions that we have reached after studying the issue.
Our main conclusion is that the earnings test should be eliminated at and above the full
benefit age. But eliminating the earnings test for those below the full benefit age would be unwise
because (1) it could increase widow poverty, (2) it is not obviously politically necessary, (3) it
involves additional short-run budgetary costs, and (4) there is no evidence that it will produce a
significant increase in labor supply.
Increased widow poverty
Eliminating the earnings test below the full benefit age could result in increased widow
poverty. It would thus pose additional challenges for ensuring that the overall Social
Security reform package reduces such poverty.
The earnings test results in higher Social Security benefits for survivors and thus provides
some protection against poverty among widows. The earnings test results in such widow
protection in two ways:
-- Directly, through the subsequent benefit increase for working beneficiaries who have
their benefits reduced under the earnings test; and
-- Indirectly, through the effect of the earnings test on delaying when beneficiaries claim
initial benefits.
The indirect effect is likely to be more important. Many analysts believe that the earnings
test is an important reason why 40 percent of beneficiaries delay initial receipt beyond age
62. Elimination of the earnings test below the full benefit age could therefore induce
increased early election of benefits. And increased early benefit election would reduce
subsequent survivor benefits (in addition to annual retiree and spousal benefits), thus
exacerbating widow poverty.¹
The effect of additional early claiming on widow poverty could be substantial: average
benefits for widows who have had their benefits reduced for early retirement are
significantly below the poverty line, whereas average benefits for widows whose benefits
are not reduced for early retirement are significantly above the poverty line. More
specifically:
-- Among non-disabled widows and widowers aged 80 and over in 1996 with benefits
reduced for early retirement, the mean annual benefit was $7,009, or $516 below the
poverty line of $7,525 for a single elderly person.
-- Among other beneficiaries (those whose deceased spouses did not claim early retired
worker benefits, and who did not claim early survivor benefits themselves), the average
benefit was $9,174,or $1,649 above the poverty line.
Differences in Primary Insurance Amounts explain only about half of the mean difference
across these groups. These figures are suggestive rather than conclusive, but we find them
troubling nonetheless.
In 1996, a significant portion of those who delayed receiving benefits until the full benefit
age or thereafter had relatively low Primary Insurance Amounts.² If the removal of the
earnings test below the full benefit age induced even a modest share of such retirees to
elect benefits earlier, the adverse consequences in very old age (either for the workers, or
their spouses, or both) could be significant. Again, further research is required on this
important topic, but the potential magnitude of the effect is sufficiently large to warrant
concern.
1
Beneficiaries could save enough of their early benefits to ensure that income in very old age remained
constant But we doubt that such saving is likely to occur; recent research strongly suggests that higher Social
Security benefits will simply lead to higher consumption levels.
2 For example, one-third of new retirees electing benefits at or after age 65 in 1996 had Primary Insurance
Amounts below $650 per month, and a similar proportion had benefits below that level. Social Security Bulletin,
Annual Statistical Supplement 1997. Table 6.B3, page 263, and Table 6.B4. page 264. Some of these beneficiaries
are officials formerly employed by state and local governments or the Federal government who were not covered by
Social Security for at least part of their careers and have other pension income for those years.
2
Politically unnecessary
Complete elimination of the earnings test (above and below the full benefit age) may
appear to have political advantages. But numerous proposals with substantial Republican
bipartisan support on the Hill only eliminate the earnings test at the full benefit age, not
below it. They thus provide a basis for bipartisan support for retaining the earnings test
below the full benefit age, while removing it at and above the full benefit age.
Such proposals include H.R. 5, the Republican leadership bill to eliminate the earnings
test, which was introduced by Rep. Sam Johnson, 83 Republican co-sponsors (including
several members of the Ways and Means Committee), and 16 Democratic co-sponsors.
Proposals to "eliminate the earnings test" that retain the test below the full benefit age also
include S. 279, introduced by Senator John McCain and 10 other Republican senators,³
and the National Commission on Retirement Policy plan introduced in varying forms by
Reps. Stenholm and Kolbe and Senators Breaux and Gregg.
Additional short-term costs
While eliminating the earnings test has roughly zero long-term actuarial cost, it does have
short-term costs. Eliminating the earnings test only at or above the full benefit age costs
about half as much in the short run about $15 billion less over 5 years -- as eliminating it
both above and below the full benefit age.
Unclear labor supply effect below the full benefit age
The traditional view of the labor supply effect of the earnings test was that it had little
impact.
A recent study for those above the full benefit age, has reached a somewhat different
conclusion, suggesting that eliminating the earnings test at age 65 would produce a 5.3
percent increase in work activity among elderly beneficiaries aged 65 and over.⁴
There is no similar evidence regarding the labor supply effect below the full benefit age.
In other words, unlike the questions the recent study raises about the traditional view of
the labor supply effect above the full benefit age, we are not aware of any credible
evidence suggesting that the labor supply effect for those below the full benefit age is
significant.
3
The bill also has one Democratic co-sponsor.
4
Leora Friedberg, "The Social Security Earnings Test and Labor Supply of Older Men," in James Poterba, ed.,
Tax Policy and the Economy (MIT Press: Cambridge, 1998).
3
Eliminating the earnings test at and above the full benefit age
Despite our arguments against removing the earnings test below the full benefit age, there
is a much stronger case to be made for removing it at that age. (A brief appendix discusses
whether it should be removed at the full benefit age, at age 65, or at age 67, which differ because
of the scheduled increase in the full benefit age.)
First, we have less concern about poverty among widows whose spouses waited until the
full benefit age before claiming benefits, so there is less need for protection in this regard.
Second, the actuarial adjustment that compensates individuals for the delayed benefits
receipt embodied in the earnings test is less equitable after age 65 than before.
Third, the new evidence on the labor supply disincentives of the earnings test does apply
to those in this age range.
Other policy suggestions
We have three further suggestions for reform that could complement our major policy
suggestion of removing the earnings test for those at or above the full benefit age, but not below
it:
Education: Concerns about keeping the earnings test in place for those age 62-64 could
be alleviated by a campaign by the Social Security Administration to make older persons
aware that the earnings test is not a tax. Even popular tax guides do not mention the
subsequent benefit increase under the earnings test. A clear and concise mailing to all 61-
year olds about how the earnings test really works, with simple examples, could
substantially improve understanding of the system. Similarly, beneficiaries whose benefits
are reduced because of the earnings test should be told how much their subsequent
benefits will be increased as a result.
Spousal benefits: Benefit reductions under the earnings test are pro-rated between the
worker and the spouse. But spouses above the full benefit age do not have their
subsequent benefits increased as a result. For a couple with a spouse past the full benefit
age, this provision partially converts the earnings test from a forced savings program to a
true lifetime tax.
5
Even after the delayed retirement credit reaches 8 percent in 2005, it will still not fully compensate those who
are 68 or 69.
6
It should be noted, however. that widow benefits are increased by the delayed retirement credit. Therefore,
the spouse may receive a higher benefit as a survivor (but not as a spouse) because of the benefit reductions under
the earnings test. In other words, she only receives the higher benefit after her spouse (the worker) dies.
4
One possibility is therefore to apply the benefit reduction under the earnings test
exclusively to the worker's benefits, rather than pro-rating it between the worker's
benefits and the spouse's benefits. That approach would obviate the need to adjust the
spousal benefit subsequently, thus removing the quirk.
Medicare coverage and election of Social Security benefits: When workers register
for Medicare benefits at age 65, they are encouraged (but not required) to elect Social
Security benefits as well - even if they plan to continue working, in which case their
benefits would be reduced through the earnings test until they reach age 70 - to ensure
that they do not later forget to elect Social Security benefits. This practice should be
ended. Indeed, if anything, 65-year olds should be presented with a table showing how
much higher their Social Security benefits would be if they waited to elect initial benefits.
Encouraging those 65-year olds working full-time to elect Social Security benefits - only
to have those benefits eliminated through the earnings test - only increases frustration with
the perceived tax and exacerbates the administrative burden of operating the system.
Changing the current practice will become even more important as the full benefit age
increases above 65.
5
Appendix: Elimination at the full benefit age versus elimination at age 65 or age 67
We concluded that the earnings test should be eliminated at and above the full benefit age,
but not below it. Some analysts have expressed concerns about linking the age at which the
earnings test no longer applies to the full benefit age:
First, it would cause the earnings test to apply to older workers as the full benefit age
increases from 65 to 67 under current law;
Second, it could create a political disincentive to increasing the full benefit age further or
faster, since it would impose a "tax" on those then falling below the full benefit age.
One twist on our proposed reform is therefore to eliminate the earnings test at age 65 so
that the age at which the earnings test no longer applies is not linked to increases (currently
scheduled or otherwise) in the full benefit age. While this proposal merits scrutiny, our initial
reaction is that it is inferior to elimination at the full benefit age itself. Since one of our primary
concerns is the impact of the earnings test removal on early claiming and therefore on widow
poverty, and since the actuarial reductions for early claiming are tied to the full benefit age, we
believe that rying the earnings test removal to the full benefit age is more appropriate than simply
eliminating it at age 65. Elimination at age 65, in other words, could encourage early claiming as
the full benefit age rises above 65. We believe that eliminating the earnings test at the full benefit
age is the most auspicious approach.
6