“The Family” in Social Security:
Entrenched Norms and Prospects for Transformation
Katina Boosalis
Introduction
Since its creation seventy years ago, the Social Security system in the United States has
grown to cover almost 53 million people with five different types of benefits.1 It is one of the
nation’s most successful programs and is responsible, in large part, for reducing the poverty rate
among the elderly. In fact, for 25 percent of elderly beneficiaries, Social Security benefits
provide 90 percent of their income.2 Although Social Security is now a massive program
covering individual workers and their dependents in the event of death, disability, or retirement,
Social Security emerged on a much smaller scale. In the beginning, retirement benefits (which
constitute the focus of this paper) were available only to a limited number of workers. 3 In 1939,
four years after the passage of the initial Social Security Act, Congress extended benefits beyond
this limited form of social insurance to cover a worker’s dependents and survivors in the event of
the worker’s death or retirement. Then and now, an individual’s Social Security retirement
payment is an “earned” benefit, indexed on the basis of the worker’s earnings and tax
contributions over the course of his or her employment. In contrast, survivors’ and dependents’
1
In August 2005, some 52,857,000 people received Social Security, Supplemental Security Income (SSI), or both.
The different forms of Social Security benefits are, in order of prevalence: retirement (old age) benefits, disability
insurance benefits, survivors’ insurance benefits, Supplemental Security Income benefits, and dependents’ (old age)
benefits. Monthly Statistical Snapshot – September 2005:
http://www.ssa.gov/policy/docs/quickfacts/stat_snapshot/index.html (last visited Nov. 8, 2005).
2
Kilolo Kijakaze, The Social Security System and Women Today: Social Security and Women of Color, 16 N.Y.L.
SCH. J. HUM. RTS 217, 225-26 (1999).
3
In addition to pension benefits, under the 1935 Act, a one-time lump-sum death benefit was available to survivors
after an insured worker died. Social Security Administration, Research Note #16: Summary of Major Benefits
Under the Social Security Program, http://www.ssa.gov/history/benefittypes.html (last visited Nov. 8, 2005).
1
benefits are available without regard to the secondary beneficiary’s participation in the
workforce and are based solely on a recognized family relationship to the insured wage earner.
In 1939 the typical American family probably stood to benefit from the conception of the
“family” that was embraced by the Social Security Act. The predominant family form consisted
of an opposite sex couple with the father in the workforce and the mother laboring for the
household at home. Today, the picture of the American family has dramatically changed.
Increased rates of divorce, decreased rates of childbearing, an increase in the number of dualearner families, and greater numbers of people living outside of marriage has greatly diversified
the picture of the American family. Although in certain ways the Social Security system has
never served all families equally, today the inequalities that result from rewarding certain marital
choices over others are more acute. In the 1930s, “nontraditional” families (i.e., families that
diverge from the two-parent, single-earner model) were likely invisible to policymakers. Today,
the fact that eligibility standards for dependents’ and survivors’ benefits remain contingent upon
marriage creates incentives for people to follow the traditional family model and disfavors
cohabitation and same-sex relationships.
Despite the program’s seeming inability to accommodate new family forms, Social
Security has in fact changed since it was first enacted. For example, in the 1970s the Supreme
Court responded to de jure distinctions within the Social Security program on the basis of gender
and illegitimacy and overturned sections of the Social Security Act that violated equal
protection.4 In addition to these judicial changes, Congress has responded to certain trends and
has enacted amendments to better accommodate, for example, divorced spouses.5 Although later
4
See infra text accompanying notes 107-148 for a discussion of these cases.
See infra text accompanying notes 79-90 for a brief history of these amendments. Also note that,
uncharacteristically, stepfamilies and adoptive parents and children were included in the Social Security Act of
1935. For a discussion of stepchild and adopted child coverage see infra text accompanying notes 91-106.
5
2
sections will explore the impact of these changes, this paper will argue that a piecemeal approach
to modernizing Social Security has failed to effectively respond to changing societal norms. In
turn, the paper will propose some modifications for the program to better serve all American
families.
Specifically, this paper will argue that benefits that are contingent on “family eligibility”
accommodate traditional families while failing to reach a growing number of American families.
These de facto features of the program reveal an implicit (or explicit) preference for the
traditional, male-headed, single-earner family. Although this preferential structure is apparent
today, the New Deal Era framers’ placement of this bias in the Social Security Act was
uncontroversial at the time. Instead, the initial Act was likely modeled on what the majority of
American families looked like in the 1930s and the Act probably also reflected a normative
conception of what the framers thought the American family should look like.
Ideally, to better equip Social Security to assist all families that face death, disability, or
retirement, policymakers would make the program more supportive of all families by decreasing
the reliance on marriage as the marker for eligibility. One suggestion for partial reform in terms
of marital eligibility is “earnings sharing,” which would equally divide a household’s total
earnings between spouses. Another proposal that would significantly impact “the family” is
privatization, which might give individuals the freedom to name their own beneficiaries.
Beginning with an overview of the formation of the Social Security Act, Part I will
contextualize the passage of the Social Security Act and delineate the significant changes that
occurred with amendments to the Act in 1939. Part I will also briefly analyze the Act in light of
the way the American family looked in the 1930s. In Part II, I will examine the ways the
program has changed in the courts and through the political process since its early years to
3
accommodate the changing American family. Part III will analyze suggestions for policy reform
and put forth two suggestions. One proposal will suggest that eligibility should be rewritten to
cover members of a “household” in light of the fact that “new” nontraditional families fall short
of the ideal family form and thereby receive less from Social Security. A final, more practical,
proposal will advocate for moderate legislative changes that would insure currently uncovered
parents and children without extending benefits to unmarried partners.
Part I: Creating “Social Security” – How the Act and its Initial Amendments supported
“The American Family” in the Depression Era
In 1935, many American families were overwhelmed with the devastating economic
consequences of the Great Depression. With unemployment soaring, the Federal government
devised a variety of responses to reintroduce some modicum of financial security to Americans
who were out of work and facing deep poverty.6 During the Depression, the elderly faced
especially dire circumstances. Having spent down their savings, senior citizens were not able to
provide for themselves and were forced to rely on charity, their extended families and adult
children, or public assistance, which was largely unavailable, for financial support.7 President
Roosevelt responded to these desperate economic conditions with a package of programs
designed to insure American families against the ravages of poverty.8
The Social Security Act of 1935, developed in response to these economic circumstances,
formed the basis for many social programs that are still in place today. The Act did not focus
solely on poverty in old age and it did more than simply provide retirement benefits to the
6
Between 1929 and 1933 the unemployment rate in the United States skyrocketed from about 3 percent to over 25
percent. Wikipedia, Great Depression, http://en.wikipedia.org/wiki/Great_Depression (last visited Nov. 8, 2005).
7
Stephen Sugarman, Children’s Benefits in Social Security, 65 Cornell L. Rev. 836, 838, n.14 (1979-80). Sugarman
notes (see n. 14) that “only 16 state [welfare plans for the elderly] were actually operating in 1933 and many had
long waiting lists.”
8
EDWARD D. BERKOWITZ, America’s Welfare State: From Roosevelt to Reagan 13-15 (Johns Hopkins Press 1991).
4
elderly. When initially enacted, the Social Security Act created four major programs which
provided cash benefits to retired workers as a type of pension (Title II), means-tested cash
benefits for poor elderly individuals (Title I), unemployment benefits (Title III and IX), and
benefits for poor, single mothers with dependent children (Title IV).9
Old Age Insurance (Title II), the focus of this paper, was designed to “provide lifetime
pensions for elderly retired workers… as a matter of right, regardless of individual need.” 10 By
deducting Social Security taxes from employees’ paychecks and by requiring equal contributions
from employers, the 1935 Act imposed a type of “forced savings” on American workers. 11 With
the Amendments of 1939, this forced savings feature began to function as a “withholding tax that
takes money from workers [who are presently in the labor force] and uses that money to pay
retirement pensions to the elderly and disabled.”12 This “pay as you go” system “links the
present state of employment to [a] future state of dependency.”13
Since Social Security
retirement payments are tied to a worker’s own labor and input into the system, to a certain
extent, individual beneficiaries pay for their own pensions.14 This contributory feature, entitles
qualified beneficiaries (i.e., any covered worker at the statutory age of retirement) to receive
Social Security retirement benefits as a matter of right.15
While the passage of the 1935 Act established the plan for distributing these retirement
benefits and death benefits (which were then a return on a worker’s contributions to the worker’s
estate), the taxation scheme to fund these benefits was not set to begin until two years later in
See Sugarman, Children’s Benefits, supra note 7 at 837 (noting that except for retirement “pension” benefits all of
the forms of assistance enacted by the Social Security Act of 1935 consisted of federal grants to states. Only Title II
retirement benefits, now known as Old Age Survivors and Disability Insurance (OASDI) was controlled in whole by
the federal government.)
10
Id. at 838.
11
Id.
12
BERKOWITZ, supra note 8, at 2.
13
Id.
14
Id. at xviii.
15
Id.
9
5
1937.16 With delayed taxation, states were to begin distributing pension checks to retirees in
1942.17 Although this five-year lapse in time between collecting taxes and disbursing benefits
was sensible from a planning and implementation standpoint, the gap between collecting the tax
and paying out benefits would lead to the accumulation of an astoundingly large surplus of 46
billion dollars.18 The fact that the federal government was to hold on to this massive sum during
a time of great need for immediate payment subjected the program to severe critique.19
Delay surrounding the actual receipt of benefits and concern about the government
surplus contributed to lukewarm support and even rabid opposition to Social Security pension
benefits leading up to the passage of the Act in 1935. In the 1930s, many politicians advocated
for more radical efforts to help people recover immediately from the effects of the Depression.
They called for emergency aid to relieve distress and advocated for “recovery [to] precede [more
long-range] reform.”20 In addition to these recommendations, which would have placed present
exigencies above the need for creating programs to prevent future economic disaster, certain
advocates pushed for much more radical social insurance plans. The chief example of this type
of proposal was the “Townsend Old Age Revolving Pension Plan.”21 The Townsend plan would
have entitled “every citizen of the United States, sixty years of age and over… to receive $200
per month. The only requirement was that the recipient [would have to] spend [that] amount
16
EDWIN E. WITTE, THE DEVELOPMENT OF THE SOCIAL SECURITY ACT 147 (The Univ. of Wis. Press 1963); Social
Security Online, Agency History, Research Note #2 – The History and Development of the Lump Sum Death
Benefit, http://www.ssa.gov/history/lumpsum.html (last visited Nov. 17, 2005).
17
BERKOWITZ, supra note 8, at 41.
18
Id. This scheme was justified because under the provisions of the 1935 Act, people needed a few years of work to
build up enough credit to deserve more than a trivial benefit. Also, from an actuarial standpoint, it made sense to
build up a large surplus. This is how private pension plans and insurance companies operate when they take in
money now with an obligation to pay that money out in the future. E-mail correspondence with Professor Sugarman
Nov. 15, 2005 (on file with author).
19
Id.
20
ARTHUR J. ALTMEYER, THE FORMATIVE YEARS OF SOCIAL SECURITY 9 (The Univ. of Wis. Press 1966).
21
Id. The Townsend Plan was named after the retired doctor who lobbied for the proposal. Dr. Francis Townsend
began advocating for the plan in 1932 and the widespread movement to enact the plan carried strength through 1952.
See id. at 99, 242-43; see also BERKOWITZ, supra note 8 at 18-19. Across the nation, as many as 25 million people
signed petitions in support of the Townsend plan and 1200 clubs were formed to advance the cause.
6
within thirty days.”22 To cover the cost of providing these sizeable “Townsend” pensions, a
universal sales tax of 2 percent would have been imposed on all transactions.23 Attracted by the
idea of such a large pension, hordes of elderly Americans, dubbed “Towsendites,” became loyal
and vocal supporters of the proposal.24
Faced with the demands of their constituents to provide immediate aid or adopt the
Townsend proposal, few politicians favored of President Roosevelt’s vision for Social Security
retirement benefits. In the end, one of the major reasons for the eventual passage of the Social
Security Act of 1935 was likely the fact that Title I, which provided federal aid to bail out state
welfare programs for the elderly was placed at the forefront of this omnibus act.25 Despite
concerns about the administration of the social insurance plan for retirement benefits (Title II),
no politicians wanted to vote against the Act as a whole and place popularly demanded welfare
benefits for the elderly in jeopardy.26
Amidst the relatively unenthusiastic enactment of Social Security retirement benefits, lay
even deeper doubts about the system’s constitutionality.27 When payroll deductions began in
1937, concern about the program became more acute and the program was challenged in the
federal courts.28
In the test case, Helvering v. Davis, a shareholder challenged the
constitutionality of the Social Security Act of 1935 and sought to “restrain [Edison Electric
Illuminating Company of Boston] from making the payments and deductions called for by the
22
ALTMEYER supra note 20 at 10.
Id.
24
See WITTE supra note 16 at 95 n.56 (noting the bulk of mail that state representatives received in support of the
Townsend plan).
25
BERKOWITZ, supra note 8 at 26.
26
Id. at 27.
27
WITTE, supra note 16 at 146 (highlighting the difficulty of drafting such a complex system from scratch and
explaining the objections and doubts as to the constitutionality of the plan expressed by leading members of the
House Ways and Means Committee).
28
BERKOWITZ, supra note 8 at 41.
23
7
Act.”29 The company’s lawyers argued that paycheck deductions would produce unrest among
workers and lead to demands for increased wages.30 Corporations and corporate shareholders
would also suffer “irreparable loss.”31 On the constitutional merits, the corporation also argued
that the tax constituted a taking in violation of the 5th Amendment and that the Social Security
Act violated the 10th Amendment.32 In a broader ideological vein, the company’s counsel also
raised the critique that “aid from a paternal government may sap [the] sturdy virtues [of selfreliance and frugality] and breed a race of weaklings.”33
In response to these arguments, the attorneys for the government simply asserted that the
Social Security Act exercised the federal government’s Article 1, Section 8 power to appropriate
and tax for the general welfare.34 The Supreme Court agreed with the federal government and
upheld the Act as a rational assertion of the taxation power.35 The Court also backed this
conclusion by describing the calamity of the Depression, the need for a uniform response to
neediness in old age, and the great hope that this Act would save “men and women from the
rigors of the poor house as well as from the haunting fear that such a lot awaits them when
journey’s end is near.”36
Even after victory in the Supreme Court, the practical future of Social Security pension
benefits remained uncertain.
Popular support for additional welfare payments (old-age
assistance, Title I) was still strong and in many states, welfare payments often exceeded the
pensions that would be available through Social Security retirement benefits (Title II).37
29
301 U.S. 619, 637 (1937); see also BERKOWITZ, supra note 8 at 42-43.
301 U.S. at 637.
31
Id.
32
See Brief for the Respondents, Helvering v. Davis supra note 29 at *3 (available on Lexis-Nexis).
33
Id. at 644-45.
34
Brief for the Petitioners, Helvering v. Davis supra note 29 at *2 (available on Lexis-Nexis).
35
Helvering, 301 U.S. at 644-45.
36
Id. at 641-42, 644.
37
BERKOWITZ, supra note 8 at 44-46, 56.
30
8
Pressure to speed up payment of benefits, to make the program a viable alternative to welfare,
and to address the question of what to do with the massive reserve forced policymakers to
consider changes to the 1935 Act.38
A. The Formation of the 1939 Amendments
To respond to critiques of the program Arthur Altmeyer, the first Chairman of the Social
Security Board, initiated a “self-conscious campaign to make the program more politically
appealing.”39 He formed a special advisory committee and began to actively defend “old age
insurance as an alternative to welfare” that “prevented dependency because it emphasized ‘thrift
and self-reliance.’”40
While pushing this rhetoric of abolishing dependency, federal
policymakers paradoxically proposed the creation of dependents’ and survivors’ benefits.41
Dependents’ benefits were not a wholly new idea. The need for such benefits was apparent since
(proposed but still undistributed) Social Security pensions to a retiree and his wife would often
times be lower than the old-age welfare benefits available in many states.42 Moreover, most old
age insurance programs in other nations already provided survivors’ benefits.43 Nevertheless,
while drafting the 1935 Act, policymakers hesitated to implement a supplementary system of
survivors’ insurance because of the fear that increasing its cost would engender greater hostility
from Congress and the Supreme Court.44 Unexpectedly, just four years after passing the initial
Act, extending benefits on a “family concept” to more individuals and increasing the cost of the
program seemed to be the best way to ignite support for Social Security insurance benefits.
38
Id. at 46.
Id.
40
Id. at 47.
41
Id. at 47-48,
42
See BERKOWITZ, supra note 8 at 56-57 (noting how, especially in rural areas, welfare benefits often exceeded the
benefits that were available from Social Security’s Old-Age insurance. Unfortunately, this discrepancy persisted
even past the initiation of the 1939 Amendments and was only addressed in the Amendments of 1950 which
expanded the program to cover more people who were employed in primarily rural jobs).
43
See Sugarman, supra note 7 at 843, n.32
44
Id.
39
9
Creating new classes of family-based beneficiaries would increase the cost of the program but it
would also reduce the surplus resulting from the tax on employers and employees. To address
concerns about delayed payment, policymakers also proposed speeding up the payment schedule
so benefits would be distributed almost immediately.45
These recommendations were enacted in 1939 and did extend benefits to a large, new
class of dependents and survivors. The 1939 Amendments to Social Security probably secured
the program’s survival by appeasing political foes since it offered higher benefits to a larger class
of beneficiaries.46 Nonetheless, by expanding the number of beneficiaries, the Social Security
system took a drastic step away from the social insurance principle that each covered worker
earned his retirement pension based on the tax he paid in while working. With the 1939
Amendments, new family benefits were extended to retired workers’ dependents and to survivors
of workers who died before retirement regardless of whether or not these family members had
directly contributed to any part of the pension.
By providing benefits to survivors and
dependents, paternalism in the Social Security system and a resemblance to welfare actually
increased since the federal government stepped in to take the place of the family’s wage earner
(i.e., father). Instead of inspiring self-reliance, the availability of benefits without a means test
encouraged dependency on the federal dole. To implement this newly family-focused program,
the federal government had to make number of (paternalistic?) assumptions about how families
functioned.47
First, the effect of adding of dependents’ and survivors’ benefits ensured that married
workers and workers with children would get more benefits than a single covered worker. 48
45
BERKOWITZ, supra note 8 at 47-48; see also ALTMEYER, supra note 20 at 90-92.
See BERKOWITZ, supra note 8 at 191.
47
Id. at 48.
48
Id.
46
10
Under the 1939 Amendments, all payments to dependents and survivors were still based on a
worker’s average monthly wage, which translated into the insured worker’s “primary insurance
benefit.”49 An individual worker would receive his primary benefit and his aged wife, dependent
children, and any eligible survivors would receive a percentage of that primary benefit. 50 For
instance, a dependent wife over the age of 65 would receive an extra fifty percent of the primary
benefit while her retired spouse received his benefit in full. By virtue of this rule, a couple
would receive 150 percent of what a single worker who earned the same average wage would
receive. Despite having contributed the same amount in taxes, the married couple received a
windfall “at no extra cost” while the single worker essentially subsidized the increased benefits
for couples.51 In addition to the benefits provided to wives over the age of 65, children of
insured retirees under the age of 18 also qualified for a benefit equivalent to half of the primary
insurance benefit.52
In addition to this inherent bias in favor of married couples, the creation of family
benefits also exposed generalizations and assumptions about gender roles and the typical
American family of the 1930s. For example, a preference for adherence to traditional gender
roles was expressed in the structure of survivors’ benefits.
Under the 1939 Amendments,
widows received 75 percent of their spouse’s primary insurance benefit. 53 The practice of giving
women only three-quarters of what a single man would have received was justified by the fact
49
ALTMEYER, supra note 20 at 101.
Sugarman, supra note 7 at 848-49.
51
Id. at 849.
52
These family maximums still apply and cut-off funds to large families by imposing caps on the amount of benefits
a family can receive. See Sugarman, supra note 7 at 869-71, 882-83. This limitation contrasts with programs like
food stamps which are based on economies of scale and do not foreclose benefits to large families.
53
See Sugarman, supra note 7 at 848, 856 (also noting that by awarding a widow with 75% of the husband’s benefit
the widow received half of what the couple would have received together (150%) while the husband was alive.
While this is true, if a wife died first leaving her husband as a widower, he would have retained his full 100%). In
1972, Congress responded to this discrepancy and increased the widow’s benefit to 100% of her late husband’s
benefit. Grace Ganz Blumberg, Adult Derivative Benefits in Social Security, 32 STAN. L. REV. 233, 241 n.32 (1980).
50
11
that women “needed less than [men].” A member of the Social Security advisory council
supported this conclusion by pointing to a women’s presumptive experience with household
work. Men needed higher benefits than women because a widow was “used to doing her own
housework whereas the single man [had] to go to a restaurant.”54 Survivors’ benefits also ended
upon remarriage.55 This illustrates how Congress assumed, and still assumes, that spouses will
take responsibility for their partners.
For widowed mothers with children, the 1939 Amendments also contained preferences
about how families with absent fathers should operate. Before the establishment of survivors’
benefits, needy widowed mothers with children had to rely on state-run Mother’s Pension
programs.56 These programs scrutinized beneficiaries closely and were often administered to
deny aid to women of color or women with objectionable morals or lifestyles. 57 In contrast to
discretionary (and often discriminatory) mother’s pensions, survivors’ benefits uniformly
extended coverage to widows of insured workers who were caring for a child under the age of
18.58 There was no requirement of economic need. Any qualified young widow received 75
percent of the worker’s primary insurance benefit and surviving children received 50 percent of
the worker’s primary insurance amount.59
For both survivors’ and dependents’ benefits,
eligibility was based solely on marriage to a covered worker and without account for actual
need.60
54
BERKOWITZ, supra note 8 at 48.
§202(d)(1)(A) of Title II of the Social Security Act under the 1939 Amendments. For a reproduced text of the
1939 Act see Public-No. 379-76th Congress 6, http://www.ssa.gov/history/pdf/1939Act.pdf (last visited Nov. 12,
2005).
56
Id. at 96.
57
See, e.g. Sugarman, supra note 7 at 876 (briefly discussing how county welfare officials often denied aid because
they disapproved of a mother’s “dating behavior or race”).
58
Id. at 848.
59
Id. at 849; BERKOWITZ, supra note 8 at 48-49. Again, this was subject to family caps.
60
In addition to benefits for surviving spouses and children, if an insured worker died without dependents, his
surviving elderly parents could also receive benefits based on his average earnings. To receive benefits under this
55
12
By extending coverage to women who would have potentially qualified for Mother’s
Pensions, Social Security swallowed up the “worthy and deserving” widows who would have
previously received welfare. This eligibility structure ensured that widowed mothers could stay
at home and care for their children. Social Security’s eligibility standards also favored widowed
mothers over other single mothers who may have never been married, had gotten divorced, or
had been deserted by their spouse. In fact, the Social Security Act of 1935 and the 1939
Amendments entirely failed to address divorce.61 Hence, only widows and their children, with
their connection to a covered wage earner were transferred on to Social Security.
While
divorced or otherwise single women receiving mother’s pensions remained subject to a means
test and moral evaluations, similarly situated women with insured spouses received guaranteed
benefits with no eligibility markers apart from caring for children under the age of 18. 62
Although differences in treatment between various classes of beneficiaries were apparent in the
1930s, increasingly disparate treatment between recipients of Social Security dependents’
benefits and welfare recipients raises additional questions about the policies’ coherence today.
These inconsistent messages about parenting and work will be addressed in more depth in
Section III.
In summary, by basing eligibility for family benefits on marital status, the 1939 Social
Security Amendments expressed a preference (perhaps more inadvertent than deliberate) for
traditional family structures and rewarded families that followed that norm with a boost in
provision for parents, elderly parents had to prove financial dependence upon their deceased child. BETTY G.
FARRELL, FAMILY: THE MAKING OF AN IDEA, AN INSTITUTION, AND A CONTROVERSY IN AMERICAN CULTURE 147
(Westview Press 1999); see also §202(f)(1) of the 1939 Act see Public-No. 379-76th Congress 7,
http://www.ssa.gov/history/pdf/1939Act.pdf (last visited Nov. 12, 2005).These “Parent’s Insurance Benefits” are
still available to “[biological, adopted, or step] parents of an individual who died… fully insured if such parent has
attained age 62…, was receiving at least one-half of his support at the time of [their child’s] death [or disability]…
and has not married since [their child’s] death.” 42 U.S.C.S. §402(h)(2005).
61
Blumberg, supra note 53 at 257.
62
See, e.g. Stephen D. Sugarman, Reforming Welfare Through Social Security, 26 U. MICH. J.L. REFORM 817, 823827 (1993) (discussing the differential treatment of Social Security beneficiaries and welfare recipients).
13
benefits. However, concluding that the only inequalities in the early years of the system were
based on marriage would be premature.63 For primarily administrative reasons, relatively few
workers were covered under the 1935 Act and the 1939 Amendments. 64 Some of the nation’s
poorest laborers, farm workers and domestic servants, were not included under the Act because
administering a tax on their wages was thought to be too complex.65 This decision carried
significant racial and geographic consequences since many agricultural workers and domestic
servants were black men and women in the South.66 Despite the minimal coverage for many
low-wage workers, the Social Security Act did include, and still retains, a feature that “give[s]
relatively larger benefits to workers receiving low wages.”67
63
Interestingly, in light of the convergence of family norms and ideals in the expression of the Social Security Act,
the very terms of the Act still covered relatively few American workers by its very terms. One example is in the age
required for receipt of benefits. In 1940, the year that benefits were first disbursed, life expectancy for women was
68.2 years and 60.8 years for men. Since benefits were only available to retirees over the age of 65, a large number
of workers would have been precluded from ever receiving retirement pensions. Kingswood College Library,
American Cultural History, The Twentieth Century, 1940 - 1949, http://kclibrary.nhmccd.edu/decade40.html (last
visited Nov. 9, 2005).
64
When benefits were first disbursed in 1940, “only about 20 percent of older Americans were eligible to receive
them.” FARRELL, supra note 60 at 147.
65
Since farm and domestic laborers were sometimes housed and fed by their employers, their room and board was
considered as in-kind income. Policymakers thought that it would “prove administratively impossible to collect
payroll taxes from [these workers].” Moreover, “it was felt that farmers would object to being taxed for old age
insurance protection for their employees.” Self-employed workers also fell outside the scope of the Act for similar
administrative concerns. Although it may seem unfair today, excluding these workers from Social Security was
consistent with other contemporaneous “laws regulating employment conditions.” WITTE, supra note 16 at 153.
66
See e.g., Michael B. Katz, Race, Poverty & Welfare: DuBois’s Legacy for Policy, 568 ANNALS 111, 120 (2000)
(explaining how Southern influence shaped Social Security legislation to exclude industries employing primarily
African-Americans); Peggie R. Smith, Regulating Paid Household work: Class Gender, Race, and Agendas of
Reform, 48 AM. U.L. REV. 851, 879 (1999) (noting the migration of newly emancipated African-Americans to
Northern homes for work as paid domestic servants); Houghton-Mifflin, Reader’s Companion to U.S. Women’s
History – Social Security Act,
http://college.hmco.com/history/readerscomp/women/html/wm_034800_socialsecuri.htm (noting that by excluding
agricultural and domestic workers the 1935 Social Security Act made three-fifths of all African-Americans
ineligible for Social Security benefits) (last visited Nov. 16, 2005).
67
See WITTE, supra note 16 at 152 (mentioning that the Morgenthau Amendment to the 1935 Act included a
principle that gave “relatively larger benefits to workers receiving low wages). Today this favorable calculation for
low-wage workers persists since 90% of the first $627 of a worker’s average monthly earnings is counted towards
the retirement benefit but only 32% of the earnings between $627 and $3,779 and only 15% of earnings over $3,779
count towards a monthly retirement benefit. Your Retirement Benefit: How it is Figured, SSA Publication No. 0510070, available at http://www.ssa.gov/pubs/10070.html#estimate (last visited Nov. 12, 2005). Maximum earnings
caps also apply which limit the amount of Social Security benefits that wealthy workers can receive. Id.
14
Given the foregoing, it is clear that from its outset, the Social Security retirement benefits
program favored certain workers and families at the expense of others. What appear today as
glaring inequalities or evidence of social engineering were possibly more justified in the 1930s.
On the one hand, the hugely daunting task of implementing an administrative system on a
national scale required clear rules with simple eligibility guidelines. Perhaps marriage was
simply an easy line to draw and participation in industrial labor markets instead of less organized
farm or domestic labor was easier for the Social Security Administration to monitor. On the
other hand, the picture of the typical American family in the 1930s likely coincided quite closely
with the benefits preferences.
Regardless of the underlying reasons, it is clear that the
policymakers who wrote the 1935 Act and the 1939 Amendments “regarded men as natural labor
force participants and women as natural child care providers” and they drafted the laws to reflect
this sexual division of labor.68
B. ”The American Family” During the First Years of Social Security’s Implementation
The reality of the American family during the 1930s, while more “traditional” than
families today, did not map neatly onto the picture of a two-parent nuclear family, supported by a
breadwinner father with a mother tending to the household and children. During the Great
Depression, rates of marriage plummeted because it became financially difficult for young
couples to form new households. In addition to a decreasing rate of marriage, the average age at
first marriage also rose during the 1930s.69 At the same time, rates of divorce declined, in part
because of the prohibitive cost of obtaining a divorce.70 Some scholars have also suggested that
68
BERKOWITZ, supra note 8 at 94.
“During the Great Depression marriage rates plummeted, as it became economically difficult for young people to
form new households. The marriage rate dropped almost 13 percent between 1930 and 1932, and by the end of the
decade the average age at marriage had risen from 24.3 to 26.7 for men and from 21.3 to 23.3 for women.”
Houghton Mifflin, The Reader’s Companion to American History – Marriage,
http://college.hmco.com/history/readerscomp/rcah/html/ah_056600_marriage.htm (last visited Nov. 9, 2005).
70
FARRELL, supra note 60 at 121.
69
15
the economic hardship of the Depression was accompanied by increased family resilience.71
Whether or not hardship actually strengthened family ties, during the Depression and certainly in
the following post-War decades, marriage was an expected life step for many Americans.
Within marriage, traditional gender roles were typically exalted.
During the 1930s and 1940s relatively few married women worked permanently outside
the home. Even though women went to work during the war effort in the 1940s, preferences for
traditional (gendered) divisions of labor persisted. For example, federally established daycare
centers were only opened with the caveat that they would close at the conclusion of the war. 72
Also, although the number of married women in the workforce increased by 50 percent from
1930 to 1940 a substantial part of this increase was probably inspired by economic need and
ended at the close of the war.73 Moreover, many of the women who found employment during
this period worked in domestic and personal services. None of these women would have been
covered under the Social Security Act.
Since ostensibly few working women were earning sufficient wages on their own and
most people believed that a woman’s proper place was in the home, the development of
survivors’ and dependents’ benefits seems logically consistent. After all, these benefits were
extended to a wife even if she never paid taxes out of her paycheck. This coincides with the
expectation that a wife would stay at home while her husband worked. Additionally, the
availability of dependents’ benefits enabled women to stay at home with children instead of
71
Id. at 108.
Id. at 44.
73
Loyola University of New Orleans, Mickey Moran, 1930s, America - Feminist Void?
The status of the Equal Rights Movement during the Great Depression, http://www.loyno.edu/history/journal/19889/moran.htm (last visited Nov. 9, 2005).
72
16
being forced to join the workforce to earn income. Again, this supported the common belief that
“helping widows stay home with their children was an idea beyond controversy.”74
This Section has illustrated how Social Security was formed and how it became
simultaneously the nation’s most expansive social insurance program and a program that
consciously wrote out certain families to favor particular family forms over others. In addition to
this marriage-based favoritism, by only reaching families with a secure attachment to the labor
force, Social Security largely failed to reach the poorest American families.75 The next Section of
this paper will explore how Social Security has changed over time to accommodate the changing
picture of the American Family. Instead of focusing on the poor families that do not qualify for
Social Security retirement benefits in the first place, the remainder of this paper will focus on the
families who are effectively denied full coverage because of their family forms and in spite of
their input from taxes and participation in the labor force.
PART II: Changes to Social Security since 1939
Over the past sixty-plus years, the Social Security program has undergone many changes.
The program has expanded to include Disability Benefits and Supplemental Security Income, a
form of means-tested benefits for very needy elderly and/or disabled individuals.76 In addition to
these expansions, Social Security retirement, survivors’, and dependents’ benefits have also
74
BERKOWITZ, supra note 8 at xvii.
Id. at 6. In addition to failing to adequately provide for people with extremely sporadic or low-wage records,
Social Security also requires that recipients are citizens or legal permanent residents of the United States. See Public
Law 104-193, the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. Certainly
undocumented workers whose paychecks are not taxed would not qualify for Social Security and these rules have
implications for immigrant families generally. However, the United States has recognized the problems that
international workers face when they pay Social Security taxes on their work in the US and in their home country.
The United States has entered into agreements with several (primarily European countries) to avoid the problem of
dual coverage and to assign Social Security coverage to the country where the worker has the greatest attachment.
Social Security Online, U.S. International Social Security Agreements,
http://www.ssa.gov/international/agreements_overview.html (last visited, Nov. 12, 2005).
76
Disability Benefits or SSDI, was created in 1956 and SSI was enacted in 1972 but brought under the control of the
Social Security Administration in 1974. SSI is an extremely important program for poor elderly couples and
individuals and SSDI assists disabled Americans of all ages. See BERKOWITZ, supra note 8 at 133; MARTHA
DERTHICK, POLICYMAKING FOR SOCIAL SECURITY 431-33 (Brookings Inst. Press 1979).
75
17
transformed in a variety of ways since 1939. Broadly, these changes fall into two types: (1)
Congressional revisions and amendments to the Social Security Act in response to changing
societal norms and (2) judicially-imposed changes in response to litigation that challenged
gender inequality and the treatment of illegitimate children in the Act. While incremental
changes have occurred within the Social Security system because of judicial and congressional
responses, overall, the program operates in much the same way as it did when benefits were first
distributed in 1940. Family benefits are still contingent on a recognized family relationship to a
covered wage-earner and the tax scheme to fund the program still functions in, more or less, the
same way.
A. The Congressional Response to Divorce
Commentators on Social Security agree that the legislation was enacted under the
presumption that families consist of “one male wage earner and one female homemaker and that
if the woman is gainfully employed her earnings contribute to her support only.”77 Notions of
“dependency” on a gendered basis and gendered “norms of labor-force participation,” have
largely persisted in Social Security despite some neutralizing trends in society as a whole. In
spite of changes in marital patterns and in workforce participation, Congress has legislated in a
piecemeal fashion to bend Social Security to respond to larger societal trends. This disjointed
approach to Social Security reform has resulted not just in inconsistencies and irrational
distinctions between recipients but has prevented productive responses to significant changes in
family life.78 Where Congress has responded to changes in the American family, it has done so
largely in the context of divorce. Because of this paper’s limited scope, this section will focus on
Congressional amendments that enhanced coverage of divorced spouses. Briefly, this section
77
78
Blumberg, supra note 53 at 243; DERTHICK, supra note 76 at 260-61.
Blumberg, supra note 53 at 257.
18
will also contrast the approach to divorce with longstanding children’s eligibility (in the context
of adopted and stepchildren) for survivors’ and dependents’ benefits.
i.
Extending Social Security Coverage to Divorced Spouses: a Tentative
Move in the Right Direction
Around the time Social Security family benefits were enacted, the “divorce rate per 1,000
married women 14 to 44 years old was 14 in 1939-41.” By 1977, that rate had risen, almost
threefold, to 38 per 1,000.79 Today, an oft-cited statistic claims that 50% of American marriages
end in divorce. While this statistic is inflated and reflects the public panic regarding the state of
marriage, the rate of divorce is still higher than it was in the 1930s.80 Despite the divorce rate,
marriage is still an immensely popular institution in the United States and fully 69% of American
women and 62% of American men marry before reaching age 35.
With nearly 95% of
Americans marrying at some point in their lifetime, the statistical prevalence of divorce indicates
that many adults (and children) will experience divorce.81
Divorce is hugely disruptive to many families for economic and personal reasons.
Women, in particular, are often adversely affected by divorce. After divorce, women are more
likely than men to assume primary care of children. Although some women may desire this
caregiver position, taking on most of the responsibility for children often leads to a significant
decline in divorced mothers’ standard of living.82 For women who adhere to traditional family
norms and work as homemakers during marriage, divorce can be especially disruptive. Women,
79
Id. n.108.
Since the 1980s divorce rates in America have actually declined somewhat. In fact, the 50% figure was always
exaggerated. See, e.g., Ira M. Ellman, Divorce Rates, Marriage Rates, and the Problematic Persistence of
Traditional Marital Roles, 34 FAM. L.Q. 1, 7-9 (2000) (citing evidence that the divorce rates declined nationwide by
about 15% from 1981 to 2000); e-mail correspondence with Ira Ellman, Nov. 21, 2005 (on file with author).
81
FARRELL, supra note 60 at 94 (citing 1997 Census figures).
82
See, e.g., JUDITH AREEN, FAMILY LAW CASES & MATERIALS 761 (Aspen, 4th Ed. 1999) (citing statistics to show
that wives experience, on average, a 30% decline in living standard after divorce but husbands experience a 10%
increase, note that these figures about the economic situation of women after divorce may ameliorate over time);
JOAN WILLIAMS, UNBENDING GENDER: WHY FAMILY & WORK CONFLICT & WHAT TO DO ABOUT IT 3, 8, 57, 115
(Oxford Univ. Press 2001)(connecting the fact that women get custody of children in 90% of divorces, are typically
cut off from the majority of the family’s wages, and experience downward mobility after divorce to the extent that
40% of divorced mothers live in poverty).
80
19
more so than men, experience inequalities stemming from caretaking and childbearing
interruptions to their careers. Women also generally earn less than men.83 Because of these
intersecting factors, divorced women often face particularly great barriers to economic selfsufficiency.
In 1950, Congress first responded to divorce in the Social Security context by creating
benefits for divorced mothers who were caring for children of their deceased, former husbands.84
Like the benefits extended to widows with children in their care, these benefits also terminated
when the youngest child reached age 18.85 This was probably adopted to take the place of
alimony and child support, which would have ended when the ex-husband/father passed away.
While creating these benefits was certainly a positive move for children from divorced families
and for their mothers, unlike benefits for young widows, benefits for divorced caregivers did not
resume once the divorced parent reached the statutory age of retirement. 86 For divorced women
who took their entitlement to mother’s benefits and postponed work until their children were
grown, the subsequent denial of benefits at retirement would have been harsh. With so many
years out of the workforce, building a sufficient individual work record for coverage under their
own Social Security account might have been difficult. Ironically, by supporting and even
galvanizing the choice to stay at home, Social Security benefits for widowed wives (divorced or
83
Blumberg, supra note 53 at 244.
In addition to expanding coverage to include divorced women, the 1950 Amendments dramatically extended
coverage to the self-employed and farm and domestic workers. The few employee classes that remained uncovered
after 1950 were religious workers and certain professional groups (lawyers, doctors, and engineers). DERTHICK,
supra note 76 at 430. Four years later, in 1954 Amendments brought more self-employed professional groups under
coverage and extended Social Security coverage on a voluntary basis to religious workers. Id. at 431. By 1965,
doctors, lawyers, and other miscellaneous professional groups were covered under the Act and today, only very few
workers escape Social Security’s wide-ranging coverage. Id.
85
Blumberg, supra note 53 at 241, now codified at 42 USCSC §402(g)(2005), “Mother’s and Father’s insurance
benefits.”
86
Id.
84
20
not) may have contributed to greater poverty later in life. On the other hand, widowed and or
divorced women who remarried would gain coverage from their new spouse’s account.
Fifteen years later, in 1965, Congress introduced a wife’s retirement benefit for divorced
wives. This amendment created more parity between divorced and widowed wives since, at
retirement age; divorced women could then receive retirement benefits based on their exhusband’s individual benefit.
Divorced wives, however, were only eligible for old-age
retirement benefits on their ex-husband’s account if they met the following three conditions: (1)
marriage for at least 20 years to her covered ex-husband on whose account she was claiming
benefits; (2) no subsequent remarriage; and (3) the receipt of actual support or the legal right to
claim spousal support from her ex-husband.87
In 1972, Congress abolished the support
requirement and a 1977 amendment reduced the threshold duration of marriage requirement from
20 to 10 years.
88
In 1977, to accord with equal protection, a court decision also extended
benefits to divorced husbands and current legislation reflects that gender-neutrality.89
The fact that Social Security now covers divorced spouses is undoubtedly an
improvement for the thousands of American families that experience economic insecurity
because of divorce. However, the limitations on benefits for divorced spouses may leave many
parents and children uncovered if the marriage that forms the basis for the derivative benefits
ends early. Any durational requirement for marriage is somewhat arbitrary since families that
fall just short of the mark (i.e., marriage lasts for 9 years and 11 months) lose out but those
marriages that last just long enough (i.e., a marriage that lasts for 10 years and 2 days) would
87
Id., 42 U.S.C. §402(b)(1)
Id. But note that it appears that the ten-year marriage rule does not apply when a widowed and divorced spouse is
caring for a child who is under 16 or disabled and is the former spouse’s natural or legally adopted child who is also
receiving Social Security benefits. Divorce Headquarters, Financial, http://www.divorcehq.com/financial.html (last
visited Nov. 9, 2005).
89
Blumberg, supra note 53 at 241-42; and see 42 U.S.C.S. §402(g), “Mother’s and father’s insurance benefits.”
88
21
qualify. Time limits for qualifying marriages might also unfairly penalize victims of domestic
violence who leave a young marriage for “good cause” because of abuse. These families might
experience even more significant barriers to finding work and becoming self-sufficient since
domestic violence so often leaves victims impoverished and without resources. 90 The ten-year
limit, however, is particularly objectionable because the median duration of marriages ending in
divorce is around seven years.91
Despite these disadvantages, a durational requirement is
probably necessary unless the entire system is overhauled in favor of income sharing, which
would make benefits portable at the dissolution of a marriage. Proposals for reform in the
divorce context are included in Section III.
ii.
Coverage under Social Security for Adopted and Stepchildren
In contrast to the lagging Congressional response to divorce, the Social Security Act has
“expressly include[d] stepchildren in the category of relatives eligible for certain family benefits”
since they were first enacted in 1939.92 The theory behind providing benefits to stepchildren is
probably an assumption that children do in fact rely upon their stepparent’s wages when their
biological parent remarries. Thus, if their stepparent were to die or retire, stepchildren have the
same need for support from Social Security as biological children. However, by providing
dependents’ benefits for stepchildren, the Social Security system seems to assume that such
children are not getting support from their biological father. Perhaps this reflects an old norm
whereby most stepfamilies were created after the death of one parent instead of after divorce. If
this is the case, the provision, at least in theory, also contemplates covering kids who could draw
90
Currently, it seems that the only provision within Social Security that specifically addresses the needs of domestic
violence survivors is a program through which, after a victim applies in person and submits evidence of abuse,
qualified survivors can obtain new Social Security numbers. Social Security Administration, New Numbers for
Domestic Violence Victims 1, available at http://www.ssa.gov/pubs/10093.pdf (last visited Nov. 9, 2005).
91
Goodwin Liu, Social Security and the Treatment of Marriage: Spousal Benefits, Earnings Sharing, and the
Challenge of Reform, 1999 WIS. L. REV. 1, 21 (1999).
92
Margaret M. Mahoney, Stepfamilies in the Federal Law, 48 U. PITT. L. REV. 491, 492, 496 n.18 (1987).
22
benefits as surviving children of a deceased biological parent with additional benefits from their
stepparent, should he or she die or retire.
Unlike adopted or biological children, stepchildren are subject to additional requirements
in order to prove eligibility for benefits. These added burdens relate to (1) the length of the
natural or adopted parent and stepparents’ marriage and (2) actual dependency. 93 For instance, a
marriage of nine months is required if the insured stepparent is deceased94 and the marriage
between the insured stepparent and child’s natural parent must have lasted at least one year if the
insured worker is disabled or retired.95
Dependency is also defined more stringently for
stepchildren than natural children.96 For biological children, dependency is presumed so long as
the child is not adopted by someone else.97 In contrast to this low burden and consequently
infrequent question of dependency for natural children, all stepchildren must prove that they
receive “at least one-half of his [or her] [support from the insured stepparent].”98
While Congress has normally tried to achieve administrative convenience in
promulgating Social Security regulations, these additional rules for stepchildren often make
eligibility questions quite complex and require close examination of family relationships.99
These added requirements evince a Congressional interest in extending benefits to stepchildren
See 42 U.S.C.§402(d)(4) (2005), “A child shall be deemed dependent upon his stepfather or stepmother…if, at
such time, the child was receiving at least one-half of his support from such stepfather or stepmother.” The option
for proving dependency by “living with” a stepparent was eliminated in 1996 (P.L. 104-121) so now all stepchildren
must show that they received 50% of their support or more from the stepparent upon whose record they are claiming
benefits. Social Security Bulletin, Program Legislation Enacted in Early 1996,
http://www.findarticles.com/p/articles/mi_m6524/is_n2_v59/ai_18747330 (last visited Nov. 14, 2005). See also
Social Security Administration, Who Can Get Benefits? http://www.socialsecurity.gov/kids/parent3.htm (last visited
Nov. 17, 2005).
94
Id. at 493, n.8; 42 U.S.C. §416(e)(2) (1982).
95
Id. at 497; 20 C.F.R. §404.357 (1999)
96
Id. at 497. The dependency requirement was only enacted in 1996. In 1996, Congress also enacted a law that
requires that a stepchild’s benefits end if the insured stepparent divorces the child’s parent after June 30, 1996.
Social Security Online, Kids and Families – Who Can Get Benefits?
http://www.socialsecurity.gov/kids/parent3.htm. (last visited Nov. 9, 2005).
97
42 U.S.C. §402(d)(3)(2005).
98
Id. at 497-98; 42 U.S.C.S. §402(d)(4)(2005).
99
Id. at 495.
93
23
only in cases of true dependency.100 That said additional factors that might support a finding of
dependency are routinely rejected in court proceedings in favor of administrative convenience. 101
Although these requirements put stepchildren at a disadvantage, the Supreme Court has upheld
the additional eligibility requirements for stepchildren on rational basis review since the
provisions aim to avoid awarding “claims based on marriages contracted solely to qualify for
benefits.”102
Adopted children, like stepchildren, have also been recognized under the Social Security
Act since derivative benefits were first created.103 As indicated earlier, like biological children,
adopted children are presumed to be dependent on their parents.104 Although adopted children
are generally treated with the same level of deference as natural children, if a parent adopts a
child after they begin receiving Social Security retirement or disability benefits, the Social
Security Administration will review the adoption to ensure that it was not entered into solely for
financial benefit.105 To qualify for benefits, an “after-adopted child” must either be the stepchild
of the adoptive insured parent or the insured parent must have legally adopted the child in a
court-certified proceeding.106 These requirements have sustained constitutional attack since they
100
Id. at 498.
Id. at 499. While it is true that a stepchild could potentially receive benefits on their natural parent’s account,
step kids would be forced to wait until their natural parent’s death, retirement, or onset of disability instead of
receiving benefits when their stepparent qualified – presumably at a point of need for the family.
102
Id. at 499-500.
103
See §§ 202(c)(3)(A), 209(k) (recognizing eligibility of adopted children for children’s benefits). The original text
of the 1939 law is available at Public-No. 379-76th Congress 5-6, http://www.ssa.gov/history/pdf/1939Act.pdf (last
visited Nov. 12, 2005). The legislative definition of “child” for Social Security purposes is now codified at 42
USCS §416(e)(2005).
104
Mahoney, supra note 92 at 497; 42 USCS §402(d)(3)(2005) (the support requirement doses not apply to adopted
children).
105
This heightened scrutiny does not apply only to stepparent relationships. Presumably, in order to prevent fraud,
any person who becomes a parent by adoption or marriage after receiving Social Security benefits is subject to this
same scrutiny.
106
See 42 U.S.C.S. §402(d)(8) (2005).
101
24
form the basis of a rational Congressional interest in discouraging the formation of family
relationships for economic motivations.107
Congress’ reluctant approach to recognizing the need for benefits in families affected by
divorce stands in contrast to the fact that Congress has always provided Social Security benefits
for adopted children and stepchildren. Although the existence of these benefits suggests some
inclusiveness in the program’s early years, children from blended families and adopted children
are still treated with a degree of trepidation by the Social Security program. An interest in
preventing fraud and limiting cost seems to cut against more liberal coverage of ”after-adopted”
kids and stepchildren. This concern has persisted despite the fact that in recent years the
coverage of stepchildren and adopted children has become increasingly important with the
greater number of divorces and the rise in second-parent adoptions by same-sex couples.
Although same-sex couples are not eligible for spousal benefits on their partner’s work record,
second-parent adoption is a means for same-sex parents to obtain full Social Security coverage
for their children. These almost certainly unintended consequences from covering adopted
children will receive more attention in the third section of this paper.
B. Judicially Driven Changes to Social Security Retirement and Family Benefits
Programs
i.
Cases on Gender
Beginning in the 1970s, the Supreme Court began to address sex discrimination under the 5 th
and 14th Amendments to the Constitution.108 By extending Equal Protection and Due Process
See e.g., Brehm v. Harris, 619 F.2d 1016 (3d Cir. 1980) (denying benefits upon holding that “after adopted”
children are not a class deserving more than rational basis review); Rodriguez v. Sec’y of Health, Educ. & Welfare,
644 F.2d 918 (1st Cir. 1981) (affirming a statutory prescription that treated adopted children who are also natural
children or stepchildren of their adoptive parents more favorably than those who are not); Stanton v. Weinberger,
502 F.2d 315 (10th Cir. 1974) (denying benefits to a child who was equitably adopted after her adoptive parent
retired).
107
25
doctrines to inequalities based on sex and gender, the Court developed an intermediate standard
of heightened scrutiny for evaluating whether gender-based distinctions were constitutional.109
A trio of cases that challenged particular sections in the Social Security Act on sex
discrimination grounds is often highlighted in Constitutional Law textbooks. The three cases
demonstrate how Equal Protection doctrine has been applied to transform aspects of the
distribution of Social Security family benefits.110 In addition to these key cases, Social Security
provisions are frequently interpreted in courts when the Social Security Administration or an
applicant appeals an administrative law judge’s decision about benefits.
Because of the
abundance of cases challenging regulations, these three decisions will illustrate the limited
situations in which courts have actually revised aspects of the program with regard to gender.
The first case, Weinberger v. Wiesenfeld, addressed the constitutionality of a distinction
under which a widow and her dependent children could all receive survivors’ benefits based on
her deceased husband’s wage record but a widower with children in his care was not eligible to
personally receive benefits on the basis of his deceased wife’s earning record.111 The appellee in
this case, Stephen Wiesenfeld, was widowed when his wife, the primary earner in their family,
died in childbirth.112 After his wife’s death, Mr. Wiesenfeld was left with “sole responsibility for
108
PAUL BREST ET AL., PROCESSES OF CONSTITUTIONAL DECISIONMAKING: CASES & MATERIALS 985 (Aspen Law
& Business, 4th ed. 2000).
109
In Frontiero v. Richardson, 411 U.S. 677, 688 (1973) the Court held that “like classifications based upon race,
alienage, or national origin” classifications “based upon [the immutable characteristic of sex]…must…be subjected
to [a] stricter standard of review.” Specifically, in Frontiero, the Court struck down the requirement that female
members of the Air Force prove dependency of their husbands in order to get an increased living allowance and
health benefits for their spouses. Id. at 678-79. In contrast to this requirement for servicewomen, the economic
dependency of the wives of servicemen was presumed. Id. at 678. The Court overturned this statute because the
differential treatment was found to serve no purpose other than “administrative convenience.” Id. at 690. Later
cases arrived at the constitutional standard of “intermediate scrutiny” for situations involving potential equal
protection violations on the basis of gender.
110
See Mary E. Becker, Obscuring the Struggle: Sex Discrimination, Social Security, and Stone, Seidman, Sunstein
& Tushnet’s Constitutional Law, 89 COLUM. L. REV. 264, 265-66 (1989).
111
420 U.S. 636, 637-38 (1975).
112
Id. at 639.
26
the care of their infant son.”113 He and his son applied for survivors’ benefits and his son was
approved. Mr. Wiesenfeld’s application was denied because such benefits were only available to
mothers.114
The Court overturned this gender-based classification upon a finding that it was based on
“archaic and overbroad” generalizations.115 The Court also cited the fact that, in 1971, 41.5% of
married women were working and that this gender-based dependency presumption denigrated
the often-significant contributions of female wage earners to their families’ support.116
Throughout the opinion, the Court noted that the purpose of survivors’ benefits for widows with
children was “to provide children deprived of one parent with the opportunity for the personal
attention of the other.”117 The Court cited the Congressional record to emphasize that the
provision focused, not on providing for dependent widows, but on permitting parents, whether
economically needy or self-sufficient, to forgo work and care for their children.118 In the end,
Mr. Wiesenfeld’s victory in the Supreme Court abolished the requirement that men must prove
dependency to recover survivor’s benefits. This equalized access to survivors’ benefits on the
basis of a marital relationship and without regard to gender.
As in Wiesenfeld, one year later in Califano v. Goldfarb, the Court struck down another
Social Security regulation that distinguished between men and women when determining
eligibility for survivors’ benefits.119 In 1950, Congressional amendments made men eligible for
survivors’ and dependents’ benefits. While all wives were presumed to be dependent on their
husbands, men who applied for derivative benefits were required to demonstrate economic
113
Id.
Id.
115
Id. at 643.
116
Id. at 643, 645.
117
Id. at 648-49 and see also the discussion about supporting parents at 648-53.
118
Id. at 648.
119
430 U.S. 199, 201-02 (1976). See also discussion of this case in BREST supra note 108 at 1008-10.
114
27
dependency on their wives.120 In Goldfarb, a widower who was denied survivors’ benefits
challenged this provision requiring a showing of dependency. Mr. Goldfarb had been in a dualearner marriage and was denied benefits after his wife’s death because he did not receive at least
one-half of his support from his wife.121 Ironically, Mr. Goldfarb had worked for the federal
government prior to retirement and was exempt from participation in the Social Security
system.122 By claiming an entitlement to surviving spousal benefits, he was actually claiming
that he had the same right as women to “double dip” and receive Social Security benefits in
addition to his federal pension.123 The Court ignored this irony and the fact that Mr. Goldfarb
was clearly not “dependent” on his wife in order to award him benefits and achieve greater
gender equity under the Social Security Act.
Although the opinion in Goldfarb recognized that Congress has “wide latitude to create
classifications that allocate non-contractual benefits under a social welfare program,” the Court
re-emphasized that even distinctions grounded in administrative convenience must meet
heightened Fifth Amendment scrutiny.124 Again, instead of looking at the statute in question as
discrimination in favor of surviving widows, the Court found that this provision’s unequal
presumptions about dependency unconstitutionally discriminated against wage-earning and
taxpaying women.
Under this regulation, working women received less Social Security
120
42 U.S.C. §402(c)-(e); Blumberg, supra note 53 at 241.
430 U.S. at 201; 42 U.S.C. §402(f)(1)(1970 ed. and Supp. V).
122
Becker, supra note 110 at 273; Goldfarb, 430 U.S. at 203. Today, Social Security has expanded to cover most
federal and state government workers. The Social Security Administration has also adopted “pension offset” rules,
which can reduce the amount of a “dually entitled” worker’s spousal or survivor’s Social Security benefit by a
fraction of that worker’s non-Social Security pension. See 42 U.S.C. §402(k)(5)(2005); Windfall Elimination
Provision, SSA Publication No. 05-10045 available at http://www.ssa.gov/pubs/10045.pdf (last visited Nov. 16,
2005); Government Pension Offset, SSA Publication No. 05-10007 available at http://www.ssa.gov/pubs/10007.pdf
(last visited Nov. 16, 2005).
123
Id. at 273 n.47 (noting that today virtually all “double dipping” is prohibited though it was initially permitted to
give female federal employees access to their husband’s Social Security benefits, which were often larger than their
personal federal pensions, without forfeiting access to their individual pension benefits).
124
430 U.S. at 210-11.
121
28
protection for their spouses despite paying the same amount of taxes as men.125 In Goldfarb, the
Court rejected old-fashioned gender role typing and noted that the taxes paid by covered female
workers actually contributed to “the fund out of which benefits would be paid to others [i.e., for
the dependents of similarly situated male workers].”126 Since the Court held that the purpose of
the 1939 Amendments was to “afford more adequate protection to the family as a unit,” it was,
perhaps, especially consistent to extend survivors’ benefits equally to all survivors in order to
protect all families.
Finally, in Califano v. Webster, the Supreme Court upheld a Social Security provision that
was found to redress “society’s longstanding disparate treatment of women.” 127
In 1956,
Congress redefined Social Security’s statutory retirement age, as 62 for women but 65 for
men.128 This gender-based distinction resulted in marginally higher benefits for women because
of a policy that allows all workers to disregard their five “lowest earning” years from the
calculation of their “average monthly wage” which forms the basis for a worker’s retirement
benefit and any corresponding survivors’ or dependents’ benefits.129
By creating a lower
retirement age for women, female workers could, in effect, exclude “three more lower earning
years than a similarly situated male [worker]” and women experienced a small boost in benefits
as a result.130 Although this statute discriminated on the basis of gender, the Court found that
this legislation was “deliberately enacted to compensate for particular economic disabilities
suffered by women.”131 In contrast to the outcomes in Wiesenfeld and Goldfarb, since this
“Mrs. Goldfarb was entitled to the dignity of knowing that her social security tax would contribute to their joint
welfare when the couple or one of them retired and to her husband’s welfare should she predecease him.” Id. at 204.
126
Id. at 206.
127
430 U.S. 313, 317 (1977)(citing Goldfarb, 430 U.S. at 209).
128
Id. at 318-19.
129
Id. at 314-16; 42 U.S.C. §415 (2005).
130
Id. at 316.
131
Id. at 320.
125
29
legislation benefited rather than burdened women on the basis of gender, the statute was upheld
as constitutional under the Due Process Clause of the Fifth Amendment.
The three foregoing cases represent the judicially enforced distinctions between remedial
(constitutional) distinctions on the basis of gender (Webster), and unconstitutional gender-based
distinctions that are rooted in old-fashioned notions of gender roles and appropriate gender
performance (Wiesenfeld and Goldfarb). In the cases involving survivors’ benefits, the Court
looked to the legislative intent behind the 1939 Amendments and concluded that the addition of
family benefits was not based on dependency but on the need for better family protection.132
While all of the cases recognized that Congress had broad discretion to legislate in the area of
social insurance, the Court emphasized the Constitution’s applicability to social welfare
programs and affirmed that mere administrative convenience was not enough to justify
discriminatory legislation.133 In Wiesenfeld and Goldfarb the Court also noted that requiring
proof of dependency actually increased administrative costs of the Social Security system and
contradicted stated government goals to achieve “administrative convenience” by differentiating
on the basis of gender.
Viewed in isolation, these cases might indicate that the Social Security system has responded
effectively, albeit through costly litigation, to problems of gender inequity. While these cases
are significant to the administration of the Social Security program, the changes that actually
resulted from the cases were relatively minimal. Certain provisions were altered in the direction
of gender-neutrality but the broader structure of the Social Security system was by no means
transformed. In fact, the cases themselves may obscure the reality behind the opinions, which
132
133
Goldfarb, 430 U.S. at 209; Weinberger, 420 U.S. at 648-49.
Goldfarb, 430 U.S. at 211.
30
still finds women at a disadvantage under Social Security in certain ways.134 For example, the
policies surrounding division of benefits at divorce disfavor women who are more likely to
experience adverse economic consequences from divorce.135 Also, these cases did not alter the
situation as faced by many families that do not fit the traditional marriage model.
ii.
Cases on Illegitimacy
In addition to these Equal Protection cases involving gender discrimination, the Court also
decided important cases regarding the rights of illegitimate children to recover Social Security
benefits in the 1970s. When family benefits were first enacted in 1939, illegitimate children
were acknowledged to the extent that they could receive child’s benefits if they were living with
and/or dependent upon their biological, insured father.136 This additional burden of proof for
illegitimate children came under attack in a case that reached the Supreme Court called Jimenez
v. Weinberger.137
Ramon Jimenez, the insured worker in Jimenez was separated from his wife when he began
living with Elizabeth Hernandez. Jimenez and Hernandez never married but they had three
children together. Their children lived with Jimenez their whole life and Jimenez became their
sole caretaker when their mother left the household after the birth of their third child.138 In 1968,
five years after Jimenez became entitled to disability benefits, he applied for child’s insurance
134
Becker, supra note 110 at 267-68. Becker makes this claim about the ways in which Social Security
disadvantages women with reference primarily to the fact that homemaking work (and all work performed outside
formal work structures) is largely performed by women and is not covered under Social Security. Id. at 269, Becker
also notes that since women’s wages are typically lower than men’s wages, by linking “life-time wages [to] old age
security” the system “inevitably creates a better social security system for men than women.” Id. at 279-80. While
Becker does note the fact that women are the primary recipients of family benefits, she questions the true advantage
that this provides to women because claims as a dependent are “smaller and more contingent… than worker’s
claims.” Id. at 281. Becker’s concerns are significant but it is difficult to determine whether or not Social Security,
on the whole, helps or hurts women. Perhaps the problem with persistent gender inequity in Social Security stems
not so much the structure of the Act, but the broader structures of societal gender inequity.
135
See, e.g., id. at 282; see infra note 83.
136
See text of original 1939 amendments §402(c)(3) available at Public-No. 379-76th Congress 5-6,
http://www.ssa.gov/history/pdf/1939Act.pdf (last visited Nov. 16, 2005).
137
417 U.S. 628 (1974).
138
Id. at 630.
31
benefits on behalf of his three kids.139 While his eldest daughter was approved for benefits since
she was born prior to Jimenez’ entitlement to disability benefits, the two younger children were
denied benefits on the grounds that “neither child’s paternity had been acknowledged or affirmed
through evidence of domicile and support before the onset of their father’s disability.”140
Jimenez challenged this determination because it was based on the “‘suspect classification’ of
illegitimacy.”141 The Court had previously used the Equal Protection Clause to strike down laws
that discriminated on the status of birth since illegitimacy was found to be immutable from the
perspective of a child who had no control over the circumstances of their birth and because it is a
status that “subjects… children to a stigma of inferiority.”142
Rejecting arguments that the classification was permissible because it had a “reasonable
basis,” the Court struck down differential treatment of nonmarital children as a violation of the
Equal Protection Clause of the Fifth Amendment.143 The Court based this holding in the purpose
of the Act, which was grounded in providing support for “dependents of a wage earner who has
lost his earning power.”144 The Court also recognized the dilemma faced by illegitimate kids
who were left with no other statutory method of proving dependency and their right to benefits.
In contrast to the Jimenez holding that affirmed illegitimate children’s eligibility for
children’s benefits, at least when their relationship to the insured parent was openly
acknowledged and affirmed, Califano v. Boles refused to extend mother’s benefits to a nevermarried mother of an illegitimate child.145 Norman J. Boles, an illegitimate child, was entitled to
child’s benefits when his biological father died since his dad had acknowledged his paternity
139
Id. at 631.
Id.
141
Id.
142
Id. referring to Weber v. Aetna Cas. & Sur. Co., 406 U.S. 164 (1972).
143
Id. at 631-38.
144
Id. at 633.
145
443 U.S. 282 (1979)
140
32
before his death.146 Norman’s mother, Nancy Gonzalez, also applied for mother’s insurance
benefits when Norman’s dad, her ex-boyfriend, died. Her application was denied. Reviewing
this denial, the District Court read Wiesenfeld to hold that mother’s insurance benefits are
primarily for the child and that the “pertinent discrimination in this case… was [at its root]
discrimination against illegitimate children.”147 The Supreme Court reversed this holding. A
narrow majority of the Court concluded that the purpose of the marital requirement was to “limit
the category of beneficiaries to those who actually suffer economic dislocation [i.e., who are
actually dependent] upon the death of a wage earner.”148
Although this view departed
dramatically from precedent by disregarding the strength of the “family support” rationale, the
Court held that denying mother’s insurance benefits to a woman who never married the father of
her child bore a “rational relation” to permissible Government purposes.149
The Court’s logic in Boles is in keeping with the preference for bright line eligibility rules,
such as rules based on marriage.
Had the Court awarded benefits to Nancy Gonzalez, a
potentially large number of cases from unwed mothers might have reached the courts and
required particularized determination. The Court’s decision is also consistent with the fact that
providing child’s benefits to Norman theoretically replaced the child support he received, or was
entitled to receive, while his father was alive. Unlike her child, as an unmarried mother, Ms.
Gonzalez would not have had a legal claim to the support of her ex-boyfriend. Perhaps this
decision is justified because holding otherwise would give Ms. Gonzalez an unexpected or
unjustified windfall when her ex-boyfriend died. On the other hand, the opinion’s insistence
about the program extending benefits only to truly needy parents is only partially true since even
146
Id. at 287.
Id. at 288.
148
Id. at 289.
149
Id. at 293.
147
33
independently wealthy widowed parents are entitled to receive caregiving benefits when their
insured spouse dies. Given this example of divergence from awarding benefits only to the “truly
needy” would support the caregiving rationale as the most logically consistent basis for
providing these parenting benefits.
This glimpse into the case law surrounding illegitimacy illustrates the Court’s foresight and
relatively flexible response to Social Security regulations, at least as compared to Congressional
responses to nontraditional families. As in the divorce context, the Court liberalized the rules
surrounding the eligibility of illegitimate children but did not (and could not) go so far as to
rewrite the rules to provide comprehensive coverage for unmarried parents or couples. Instead,
the Court retained marriage as the marker for eligibility for adults and created a relatively narrow
exception for non-marital children who could demonstrate actual dependency.
The remaining gaps in Social Security coverage for unmarried families, married families that
break up, and families that deviate from the traditional model will be discussed in more detail in
Section III. Although Section III will offer some suggestions for more systemic policy changes,
it is unlikely that judicial opinions will ever trail blaze in the area of Social Security reform.
Since courts must defer to other governmental branches and should respect precedent, more
revolutionary and systemic change in the area of social insurance will probably have to come
from elsewhere.
Part III: Options for Addressing Social Security’s Inadequate Coverage for Certain
Family Types
The foregoing sections of this paper have demonstrated that while Social Security’s
initial design may have suited many families in the 1930s, the program has responded slowly to
changing demographics and social trends. This section will explain how the program’s hesitant
34
liberalization has made the program less effective and supportive for today’s families.
Specifically, the program shortchanges families that are not the normative heterosexual marriage
preferred under the current Social Security system. As in many large, bureaucratic enterprises,
inertia within the Social Security program has resulted in a system that is unresponsive to change
and reflects longstanding biases against nontraditional families. Given the preference for twoparent, single breadwinner families, families that deviate from this ideal are disadvantaged.
These include dual-earner families, families that experience divorce, as well as cohabitants and
their children. After exploring the inequities that these groups face under today’s Social Security
system this section will propose policy approaches for reform.
A. Dual-Earner Couples and Divorced Families
Johnny and Gina are married with three kids. Both of them work – Gina works at a diner
earning $30,000 per year and Johnny used to work on the docks, but when his union went on
strike he became a truck driver, and now earns $30,000 per year. Their children spend their
days at school or in daycare. Gina and Johnny pay $620 a month in Social Security taxes. Ruth
and Peter live next door to Gina and Johnny. Peter is an adjunct professor at the local
university and earns $60,000 per year. Ruth volunteers, cares for their two young children, and
plans to be a homemaker, at least until their children leave home to attend college. Ruth and
Peter also pay $620 a month in Social Security taxes on Peter’s income. At retirement, Johnny
and Gina will receive less retirement income ($2027 per month) than Ruth and Peter ($2220 per
month) despite the fact that they earned the same amount as their next-door neighbors and paid
an equivalent amount of Social Security taxes. In the event that either Peter or Johnny dies in
old age, this inequity would also carry over to the women’s widow’s benefits. Ruth would
receive Peter’s full benefit, but Gina would go on receiving only her individual entitlement since
it would be equivalent to what Johnny’s was. While Ruth would experience a 33% deduction in
her standard of living when she was widowed, Gina would lose half of her income when Johnny
died. Likewise, if either husband were to die young, Ruth and her children would receive larger
survivors’ benefits than Gina and her kids. The only potential area in which Gina and Johnny
would experience an advantage over Peter and Ruth is if the couples were to get divorced or if
either wife was to become disabled or pre-decease her husband. With her own earning record,
Gina would likely be better equipped to care for herself and her children than Ruth in the case of
divorce. By virtue of her participation in the workforce, Gina, but not Ruth, would also be able
to provide Social Security benefits for her children and spouse if she died or became disabled.150
This scenario is roughly based on a combination of Goodwin Liu’s introductory hypothetical in Social Security
and the Treatment of Marriage supra note 91 at 1 and Dorothy A. Brown’s hypo in Social Security and Marriage in
Black and White infra note 158 at 119.
150
35
The situation described above outlines a longstanding bias in the Social Security system
in favor of single-earner couples. Although couples like Johnny and Gina, who are both in the
workforce, are on the rise, Social Security retains a preference for couples like Peter and Ruth.
So called, “working families,” with two employed spouses actually constitute the majority
(53.7%) of American families.151 Merely 18.7 percent of all American families are “traditional,”
in which only the husband is employed.152 In 1997, 70.7 percent of married American mothers
with children under the age of 18 were in the labor force. 153 The prevalence of working families
is not a wholly new phenomenon. For instance, it was thirty years ago, in 1975, when a majority
of American mothers with school-age children first held jobs outside of the home.154 Although
familial norms have certainly changed since Social Security’s inception, the program still
operates in favor of an increasingly uncommon traditional family form. The ways in which
Social Security family benefits prefer single-earner couples are pervasive and varied.
Under the Social Security retirement benefits plan, the spouse of an insured worker can
claim benefits on their own individual account (built from their employment) or on their spouse’s
account.155 Although any spouse is eligible for a spousal benefit156, working spouses receive
their own benefit or a spousal benefit based on 50% their spouse’s earnings, whichever is
greater.157 Spouses who do not work or who contribute less than 20% to the household income
are “dually entitled” to receive benefits as a spouse or an individual. However, such spouses will
151
Cornell Employment and Family Careers Institute, Facts about the Demographics of Working Families,
www.lifecourse.cornell.edu/archives/factsht/fsfeb99.pdf (last visited Nov. 12, 2005).
152
Id.
153
Id., see generally Linda J. Waite & Mark Nielsen, The Rise of the Dual Earner Family, 1963-1997, in WORKING
FAMILIES: THE TRANSFORMATION OF THE AMERICAN HOME 27-41 (Rosanna Hertz & Nancy L. Marshall eds., Univ.
of Cal. Press 2001).
154
BERKOWITZ, supra note 8 at 9-10.
155
Liu supra note 91 at 12.
156
Subject to disqualifications due to divorce.
157
42 U.S.C. §402(b)(2)(2005).
36
always receive a larger benefit by taking the spousal benefit over their own benefit. 158 People
like Gina, who earned the same amount as their spouse, will receive no spousal benefit at all.159
This rule “enables a one-earner couple to receive greater Social Security benefits than a twoearner couple who has the same total earnings and who pays the same amount of payroll
taxes.”160 This system rewards married couples with very little in terms of Social Security
benefits for a second worker’s taxable earnings.161
At death, as at retirement, surviving spouses in single-earner families fare better than
widows or widowers in “working families.” When one spouse dies, the surviving spouse at
retirement age is entitled to receive the greater of (1) their own individual benefit or (2) the full
amount of their deceased spouse’s Social Security benefit. 162 Because of this rule, a surviving
spouse of a one-earner couple will always receive a larger survivor’s benefit than a surviving
spouse in a dual-earner couple even if they earned the exact same amount.163 This redistributive
effect means that dual-earners are taxed above and beyond single-earner families and it is the
single-earner families who reap the greater survivors’ and retirement benefits. In other words,
when it comes to Social Security benefits, for dual-earner couples, you don’t always pay for
what you get.
Despite this account of unfairness, dual-earner families do experience some advantages
over traditional families. Sadly, these boons come only with unfortunate events like divorce,
death, or disability. First, the effect of divorce can be especially grave for spouses who do not
work. Divorced spouses (primarily women) who “stayed at home” may face greater barriers
158
Dorothy A. Brown, Social Security and Marriage in Black and White, 65 OHIO ST. L.J. 111 116-19 (2004).
Id. at 119. In fact, any spouse whose contribution to the household income “reaches slightly over 20%” will
receive no spousal benefit under the “whichever is greater rule.” Id. at 118.
160
Liu supra note 91 at 13.
161
Id. at 15; 42 U.S.C. §402(e)(2)(A)(2005)
162
Liu, supra note 91 at 16 citing 42 U.S.C. §§402(e)(2)(A), (f)(3)(A) (2005). Under this rule spouses like Gina,
whose benefit is equivalent to her spouse’s, will receive solely their own benefit at death.
163
Brown, supra note 158 at 121.
159
37
than working wives to earning sufficient income to support themselves and build an individual
Social Security account. Additionally, since entitlement as a divorced spouse is contingent on a
10 year marriage, a large number of divorced spouses will not recover any dependents’ or
survivors’ benefits at all. This 10 year minimum is especially problematic in light of the fact that
“median duration of marriages ending in divorce [was] 7.2 years” in 1997.164 With that fact in
mind, working spouses are probably better off than non-working spouses since they can always
count their working years to their eventual Social Security benefit. Likewise, a working spouse
is constantly building earnings credits that can benefit his or her family in the event of death or
disability.
Lest we conclude this discussion of inequality between single and dual earner families
too soon, it is important to recognize how the variables of gender and race further complicate the
picture. First, since men tend to earn more than women, men are more likely to be primary
earners. Women thereby constitute the vast majority of recipients of spousal benefits. 165 With
time, if the gendered wage disparity diminishes, this trend may wane but in 1996, 99% of retired
individuals receiving spousal benefits were women.166 Secondly, the fact that men are usually
the primary earners in a “working family,” by virtue of earnings disparities that favor men over
women, also means that women’s wages are usually discounted or do very little to increase a
woman’s individual retirement benefits.167 Although cases like Goldfarb insured bare gender
equity so that women received the same right to building survivors’ insurance coverage for their
dependents as men, judicial intervention did not change the underlying and unequal framework
164
Liu, supra note 91 at 21.
Id. at 12.
166
Id. The same is true for recipients of survivors’ benefits since wives often earn less than their husbands and tend
to outlive their spouses as well. Id. at 17.
167
Id. at 14.
165
38
of gender and labor. Finally, the fact that women are usually more adversely affected by divorce
than men also contributes to inequity under the system.
Inequalities between working and traditional families are also exacerbated along racial
lines. Across all income levels, black couples are more likely to be dual-earners than white
couples. Married black couples are also more likely to contribute roughly equal amounts to the
household income than white couples.168 Since roughly equal wage-earning households receive
(according to their input) the least amount in survivors’ benefits, black families tend to receive
proportionally less from their Social Security contributions than white families.169 In addition to
the different experience within marriage between black and white couples, blacks are less likely
to be married in the first place than whites.170 Without the formal sanction of marriage, a partner
is not eligible for dependents’ or survivors’ benefits. Although with proof of paternity a child of
an unmarried couple may qualify for child’s benefits in the event of disability, death, or
retirement, a long-term partner categorically cannot qualify for benefits without a connection
through marriage. Therefore, lower rates of marriage in the black community probably also
contribute to less Social Security coverage.
On the other hand, Social Security does include certain features that benefit people of
color. For instance, Social Security benefits comprise more of the income for elderly people of
color than for elderly White recipients.171 Also, Social Security’s “progressive” benefit formula
gives lower-wage workers a larger percentage of their pre-retirement earnings than workers with
higher incomes. This formula benefits people of color, especially women of color, since they
168
Brown, supra note 158 at 111, 129-30.
Id. at 124.
170
“In 1980, for example, well over half of all black births occurred outside of marriage. Most black children lived
in one-parent families.” Berkowitz, supra note 8 at 8, 142. “The federal Center for Disease Control and Prevention
reports that by the age of 30, 81 percent of white women and 77 percent of Hispanics and Asians will marry, but
only 52 percent of black women will do so.” Andrew Herrmann, Marriage rate low in the black community,
CHICAGO SUN-TIMES, June 9, 2003, at 8.
171
Kijakaze, supra note 2 at 226.
169
39
tend to earn less than the majority of the country.172 This is not a result of extra generosity, but a
reflection of relative poverty. Additionally, the existence of disability and survivors’ benefits
may especially assist families of low-wage workers who are employed in physically arduous
jobs.173 Since African-American men have shorter life expectancies than the rest of the male
population, survivors’ benefits are more crucial for their spouses, who are primarily AfricanAmerican women.174 Although all of these facts describe why Social Security may be especially
important to people of color, the facts are not a panacea. Since Social Security benefits are often
not large enough to raise a recipient out of poverty and because low-wage workers are less likely
to have independent retirement savings or access to private pensions, the system still leaves
plenty of insured individuals in poverty.175
B. Cohabitants: Problems for Partners and Children
Anne and Mariko have been in a relationship for years. Both women are raising Jamie,
Mariko’s daughter from a prior, unmarried relationship. Anne works as a chef and Mariko is an
accountant. Although these women support each other and their daughter and both pay Social
Security taxes on their earnings, their Social Security contributions will not benefit their partner
in the event of their death or disability. In addition to the unavailability of spousal benefits for
either Mariko or Anne, Jamie will not qualify as Anne’s dependent for survivors’ benefits.
Simply because of the fact that their relationship is not sanctioned by marriage, Anne, Mariko,
and Jamie are ineligible for many family Social Security benefits.
Since the 1930s, marriage, while still extremely prevalent, has become increasingly
uncommon in the United States. More couples are choosing to live together outside of marriage
and increasing numbers of same-sex couples live together without the benefit of state-sanctioned
marriage. Statistics about cohabitation are difficult to pin down since data has only recently
172
Id. and see Kilolo Kijakaze, Low-Wage Earners: Options for Improving Their Retirement Income, in THE
FUTURE OF SOCIAL INSURANCE: INCREMENTAL ACTION OR FUNDAMENTAL REFORM? 44-46 (Peter Edelman et al.
eds., Brookings Inst. Press 2002)(discussing trends among occupations and earnings of African-American and
Latino workers which indicate that the median earnings for people of color are significantly lower than for White
workers). Unfortunately, none of these studies mentions Asian workers.
173
Kijakaze, supra note 2 at 227 (noting that while 12 percent of the labor market is comprised of AfricanAmericans, 19 percent of female workers who received disability benefits in 1997 were African-American women).
174
Id.
175
Kijakaze, “Improving Income” supra note 172 at 47 – 48.
40
begun to be collected. For instance, same-sex couples were counted for the first time in the 2000
census. In that census, more than 3 million gays and lesbians identified themselves as members
of same-sex couples.176 Same-sex couples were identified in 99.3% of all U.S. counties in the
2000 census.177 Like same-sex households, the number of unmarried partner households has also
risen significantly in recent years and rose by 72% in the 1990s. 178 While for some unmarried
partners, cohabitation is a steppingstone to marriage, more and more couples are choosing to live
together outside of marriage permanently. Although more couples are choosing to live as longterm cohabitants, such relationships in the United States usually break up or end in marriage
within five years. For this reason, the exclusion of unmarried partners is primarily a problem for
same-sex couples.179
Both couples who choose to build their lives together outside of marriage and same-sex
couples who are not legally permitted to marry lose Social Security benefits that are extended to
their married peers.
Although decisions not to marry carry perilous Social Security
consequences for younger couples, senior citizens may choose to cohabitate in order to retain
Social Security or other retirement benefits.180 For example, if a widowed woman were eligible
to receive survivors’ benefits on her late husband’s account, she may choose not to marry her
176
This figure is almost certainly an undercount since the census only addressed couples who were living together
and who chose to identify their sexual orientation. Additionally, the 2000 census did not include space for
individuals to identify as transgender or bisexual so the numbers may have also missed a significant portion of the
Gay, Lesbian, Bisexual, and Transgender community. Margie Mason, Census Figures on Same-Sex Couples (Aug.
8, 2001), http://speakout.com/activism/apstories/10044-1.html (last visited Nov. 12, 2005).
177
GLTBQ – An Encyclopedia of Gay, Lesbian, Transgender, Bisexual and Queer Culture, Census 2000:
http://www.glbtq.com/social-sciences/census_2000.html (last visited Nov. 12, 2005).
178
Robin Fields, “Married with Children” Still Fading as a Model, LOS ANGELES TIMES, Aug. 20, 2001, at A1
available at http://www.contemporaryfamilies.org/public/articles/change39.htm (last visited Nov. 12, 2005).
179
E-mail correspondence with Ira Ellman, Nov. 21, 2005 (on file with author).
180
Robin Fields, They're Older but Not Old-Fashioned About Love and Marriage, LOS ANGELES TIMES, Aug. 23,
2001, at pt. 5, pg. 1, available at http://www.contemporaryfamilies.org/public/articles/change52.htm (last visited
Nov. 12, 2005). Also note that California’s domestic partner laws explicitly contemplate these decisions by
extending domestic partner benefits if one or both members of an opposite-sex couple “meet the eligibility criteria…
for [Social Security] old-age insurance benefits” or if “one or both of the [partners are] over the age of 62.” CAL.
FAM. CODE §297(b)(5)(B)(2005).
41
older partner so they can both retain higher Social Security benefits. Regardless of the reasons
for cohabitation, many unmarried couples and their children receive less in Social Security
family benefits than similarly situated married couples. Federal policies and laws that favor
marriage and define it to exclude same-sex unions have stunted and curtailed policy development
that would provide cohabitants with full Social Security coverage.
i.
Policy Proposals for Social Security Reform to Better Assist All Families
Although recognition by analysts and commentators that Social Security prefers married,
single-earner couples is common, proposals addressing unfairness to unmarried couples and
nonmarital families are not. The limited attention to the plight of unmarried families in the
Social Security context probably stems from the current legal and political barriers to their equal
recognition and their relatively recent attraction of attention and concern. Scholars have devised
a variety of proposals to create more equality between single-earner and dual-earner couples.
Yet for a number of reasons, including resistance in the political process and judicial deference
to legislative decision-making, suggestions to fully overhaul the family benefits system have not
been popular. This section will briefly describe the barriers to reform and will assess prominent
recommendations for change to bring about more equity for all families. In the end, I will posit
two ways to reform Social Security, which might achieve more equal outcomes for the
nonmarital families that currently lose out in the Social Security family benefits system and
receive little attention from other proposals for change.
1. Suggestions for Social Security Reform and Obstacles to Implementation
One type of proposal, the “family-focused” proposal, has concentrated on creating more
equality for all married couples. While still operating with marriages in mind, these plans have
also aimed to address the problems that families face at divorce. A second type of proposal for
42
reform has focused instead on the fiscal soundness of the program. Both of these proposals have
social implications but the proposals for privatization seem more removed from family policy.
Nonetheless, even these fiscally-oriented suggestions would have a significant impact on
families and the normative meaning of “the family.” In fact, privatization proposals have the
potential to essentially disregard marriage as a required eligibility factor for family benefits. As
such, privatization, in contrast to “family-focused” proposals for reform, might include
cohabitants and same-sex couples in its vision of a revamped Social Security system.
a. “Earnings Sharing” – A Family-Focused Proposal
One proposed method for Social Security reform that has received much scholarly attention
is “earnings sharing.” Earnings sharing is “family-focused” in my terminology because it aims
to equalize treatment of all families formed by marriage. Earnings sharing would “eliminate the
current Social Security system’s spouse and surviving spouse benefits. In place of today’s
system, each spouse in a married couple would be credited with one-half of the couple’s
combined earnings during marriage.”181 Effectively, the present system of basing benefits on an
earner’s average monthly income would be replaced with a basic community property model for
dividing Social Security benefits. Earnings sharing would probably improve the situation of
working (dual-earner) families as compared to traditional (single-earner) families and would
ameliorate the consequences of divorce for stay-at-home spouses and secondary earners.
Recall, for example, Johnny and Gina (the dual-earner family) and Ruth and Peter (the
single-earner family). By crediting each partner in these couples with half of their total wages,
each spouse in these sample families would be credited with $30,000 worth of earnings for every
181
Jonathan Barry Forman, Making Social Security Work for Women and Men, 16 N.Y.L. SCH. J. HUM. RTS 359,
371 (1999).
43
year the couples earned $60,000.182 This would wipe away the extra benefit for single-earner
couples and the system would finally adopt an essentially neutral stance “with respect to the way
married couples decide to share breadwinning and homemaking activities.”183 In addition to this
newfound equivalence in tax contribution and benefits received, earnings sharing would make
benefits fully portable in the event of divorce. No matter the length of marriage or subsequent
marital decisions, each spouse in a divorced couple would be entitled to their half of the
community’s Social Security contributions made during their former marriage.184 By giving
marital partners an automatic half-share of the community’s Social Security earnings, earnings
sharing would eliminate the need for the 10-year marriage rule for divorced spouses. This might
particularly benefit victims of domestic violence who would share in the community’s benefits
even if a marriage broke up before the 10-year mark.185 In spite of these advantages for divorced
spouses, under earnings sharing, protection for widows who receive survivors’ benefits could
decrease.186 For instance, whereas a surviving spouse at retirement age under the current system
receives 100% of the insured spouse’s benefit or 100% of her own benefit, a surviving spouse’s
182
For years a spouse worked outside of marriage (either prior to or after marriage) those earnings would factor
separately into their individual benefits account. Again, this would work in the same manner as the community
property system wherein property brought into the marriage by one spouse is considered the separate property of
that spouse.
183
Liu, supra note 91 at 26.
184
Blumberg, supra note 53 at 288.
185
Another change that might benefit domestic violence survivors within the confines of the current system would
be development of a test whereby a divorced spouse could prove that abuse was the reason for the early end to the
marriage and then claim benefits as though the marriage had lasted for 10 years. This proposal would increase the
amount of individualized determinations that the Social Security Administration must make each year and would
probably increase the litigation traffic over Social Security in the courts. On the other hand, fear of retaliation and
the desire to fully terminate any connection with a batterer might cut against much use of this provision. Whether or
not concerns about administration proved warranted, other busy administrative departments already make these
types of determinations (i.e., the IRS reviews Innocent Spouse petitions to be relieved from income tax
underpayment penalties) and the standard for requesting relief as a domestic violence survivor could be modeled on
such programs. For information on “Innocent Spouse Tax Relief” see Innocent Spouse Tax Relief,
www.irs.gov/pub/irs-pdf/f12507.pdf (last visited Nov. 9, 2005). Additionally, the Social Security Administration
already reviews applications for evidence of domestic violence from survivors who wish to change their Social
Security Numbers. See note 90.
186
Karen C. Burke & Grayson M.P. McCouch, Women, Fairness, and Social Security, 82 IOWA L. REV. 1209, 123233 (1997).
44
benefit could fall to one-half of the couple’s total pre-death benefit under earnings sharing. This
significant income drop could make widow’s financial circumstances considerably less secure.
Additionally, it seems possible that benefits for divorced spouses who were unable to re-enter the
workforce and build their own Social Security account might also decline since under earnings
sharing they would not reap the benefit of their spouses’ post-divorce earnings. Instead, a
divorced spouse would only be able to claim their half of what the community earned during the
duration of the marriage.
Earnings sharing would effect positive changes toward equity and adequacy for all married
couples. Nevertheless, there are significant disadvantages to the earnings sharing system and
major impediments stand in the way of its adoption. While a shift to earnings sharing would
implicitly value the household work of a spouse who stayed home, earnings sharing would not
value all household work equally. For instance, regardless of how household work was actually
divided between a couple, couples with nearly equal incomes would get essentially no “credit”
for housework.187 Housework done by partners in a household like Gina and Johnny, where the
couples earn the same salaries, would be invisible for purposes of Social Security benefits. On
the other hand, the stay-at-home wife of an executive earning $100,000 per year would have her
housework valued at a salary of $50,000.188 More importantly, earnings sharing would retain the
bias in favor of married couples. Under earnings sharing, single people would have the same
rate of return on their Social Security taxes as married couples who earned the same amount.
Because single people do not benefit from the same economies of scale as couples, this system
187
Consider the fact that women typically perform more household work (their “second shift”) than men even when
both partners in a heterosexual relationship work outside the home. See generally, ARLIE HOCHSCHILD, THE
SECOND SHIFT: WORKING PARENTS AND THE REVOLUTION AT HOME (Penguin, 2003).
188
See Liu supra note 91 at 37-38.
45
would still leave singles at a disadvantage as compared with married individuals.189 Finally, a
shift to earnings sharing would improve the Social Security income of dual-earner families but
would reduce the Social Security benefits to single-earner families. This appears to be the most
salient objection to the plan and would probably be the most significant quarter from which
objection to earnings sharing would be raised. Instead of pointing out that earnings sharing does
not go far enough, single-earner families would be able to point to losses actually incurred
because of earnings sharing. Since traditional families would lose under an earnings-sharing
regime, implementation of earnings sharing might face political opposition from traditional
families and those who favor them.
In addition to these disadvantages, earnings sharing adopts a very particular “partnership”
theory of marriage, which would conflict with established family law in certain states. Since
earnings sharing would divide benefits equally in the case of divorce, adopting an earningssharing approach would result in a type of divorce reform.190 Federal preemption doctrine may
also prevent states from dividing Social Security benefits in state divorce proceedings. 191
Finally, since earnings sharing would mandate a sort of community property for “federallycreated marital property,” the imposition of federal law in state family law cases might be met
with resistance from state judges and politicians and might violate principles of federalism.192 In
light of these problems inherent in the earnings sharing model, Professor Goodwin Liu and other
scholars have suggested adopting a more radical approach that would value “non-wage work on
its own terms.”193 Notable proposals, like Nancy Staudt’s, would tax housework so homemakers
189
Id. at 30-32.
Id. at 40.
191
Id. at 41, citing Hisquierdo v. Hisquierdo, 439 U.S. 572 (1979). On the other hand, state law, even in common
law marital property states, allows for earnings sharing when dividing private pensions. See e-mail correspondence
with Ira Ellman, Nov. 21, 2005 (on file with author).
192
Liu, supra note 91 at 43.
193
Id. at 63.
190
46
can build Social Security accounts.194 Housework could ostensibly be valued on an economic
model of replacement wages (i.e., what it would cost to purchase the services of a homemaker on
the market). Unfortunately, these proposals seem fraught with so many administrative
complexities as to be impossible to implement. Given these concerns, earnings sharing is not
likely to be enacted anytime soon.195
In contrast to these ideas, which would completely override today’s system of Social Security
family benefits, other family-focused suggestions for reform would simply alter the percentage
of a spousal benefit. For example, Robert Ball, Commissioner of Social Security from 1962-73,
suggested alleviating the bias against single persons and working couples by reducing the
spousal benefit of a nonworking spouse from 50% to one-third of the worker’s primary
benefit.196 Although spousal benefits would have dropped, benefits for a single worker would
have increased by 14 percent.197 That change in both directions would neutralize the effect of
the percentage alteration and leave married couples in the same situation. While this proposal
would have equalized the situation of working couples vis-à-vis traditional couples without
making any couples suffer loss, the cost of the proposed change probably hindered the proposal’s
adoption.198
Instead of applying percentage reductions and increases to spousal and individual benefits,
another approach for reform would be to create parity between identically situated single and
dual earner couples. Taking Gina and Johnny and Ruth and Peter as examples, this approach
would create the fiction that Gina and Johnny were in a single earner family. Under the current
system, at retirement Ruth and Peter will receive Peter’s full benefit (for ease of calculation, let’s
194
See generally, Nancy C. Staudt, Taxing Housework, 84 GEO. L.J. 1571 (1996).
See id. at 47-58 (discussing the intricacies of these proposals).
196
DERTHICK, supra note 76 at 262.
197
Id.
198
Id. at 263.
195
47
say $700) plus half of that benefit ($350) for Ruth as a spousal share for a total of $1050 per
month. Because of their lower wages, we’ll assume Gina and Johnny would each receive $450
for a total of $900 per month. If Gina and Johnny’ earnings were considered together, as though
one worker had earned $60,000 per year, they would receive the same amount as Ruth and Peter
instead of $150 less per month.199 While this proposal would be easy to implement without
applying a percentage across the board, it would also increase the cost of the program.
Moreover, it would not address the fact that single earners would still receive proportionately
less in benefits for their Social Security tax contributions.
b. Proposals for Privatization and their Unarticulated Vision of “The Family”
In 2002, President George W. Bush proposed the partial privatization of the Social Security
system. If adopted, in a privatized system individual workers would contribute part of their
Social Security taxes into individual accounts.200 Although an individual might be able to
purchase insurance on their account for a spouse or child, privatizing the system would abolish
the current system of spousal benefits. By eliminating benefits for married partners,
privatization could give individuals more freedom to choose how to allocate their benefits. In
other words, in place of federally mandated decisions about which family members qualify for
dependents’ and survivors’ benefits, individual workers might get to name their dependents.
This would open the door for same-sex couples and cohabitants to choose how best to recognize
199
The idea is to combine the incomes in a dual earner family and, using that total, add in the 50% spousal benefit so
that the couple receives retirement benefits equivalent to what a “traditional” family with the exact same earnings
receives under the current law. With the death of one spouse, this system could also work to ensure that a surviving
spouse in a dual-earner couple receives what would be equivalent to the 100% individual benefit in a (financially
equivalent) single-earner family.
200
Under partial privatization people could then invest a portion of their individual accounts in the stock market.
For a summary of the debate surrounding the privatization of Social Security see Wikipedia, Social Security Debate
(United States): http://en.wikipedia.org/wiki/Social_Security_debate_%28United_States%29 (last visited Nov. 12,
2005).
48
and accommodate their families. Affording workers this type of autonomy would also end the
current preference for single-earner, traditional families.
Although this is not its intent, privatization would make a very significant impact on spousal
benefits. There is currently no way for unmarried couples (like Mariko and Anne, for example)
to provide spousal benefits for one another. Same-sex couples are barred from marriage for all
federal purposes and opposite-sex couples are not allowed to name a designated beneficiary (or
pseudo-spouse) to receive their Social Security benefits. The rules for children differ and are
already more accommodating of unmarried relationships since children of unmarried couples can
currently qualify for children’s insurance benefits if they can prove that they qualify for benefits
as an illegitimate child or if their non-biological parent adopts them. While adoption is an option
for unmarried couples to gain coverage for their children, the cost of adoption may be prohibitive
for many families.
In any event, the potentially positive outcome that privatization would have on unmarried
couples has not been championed by privatization’s proponents. On the contrary, the effects of
privatization that would undermine “the (traditional) family” seem to have escaped the attention
of political conservatives who advocate privatization for fiscal reasons.201 Without delving into
the financial pros and cons of privatizing the system, most people are quick to recognize the
downfalls of privatization.202 In chief, privatization would do very little to assist people who
cannot save adequately for retirement and who end up impoverished in their old age.
Privatization would likely have an especially detrimental effect on low-wage workers since
contributions to individual accounts would be diminished by administration and annuitization
201
That lack of attention led me to conclude that the vision of “the family” under a privatized system is somewhat
“unarticulated.”
202
C. Eugene Steuerle, Symposium, Principles of Equity and Spouses' and Survivors' Benefits, 16 N.Y.L. SCH. J.
HUM. RTS. 217, 224 (1999).
49
costs, accounts of poorer workers would be smaller and would realize lower rates of return than
the accounts of wealthier workers, and low-wage workers generally have less ability to purchase
sound investment advice.203
Despite this negative side of privatization, it does seem like the most feasible option for
extending Social Security to all people regardless of marital status.204 While it might seem like a
positive move for cohabitants and same-sex couples, privatization would actually knock all
families down to a lower level of protection. It is important to keep that reality in mind since
privatization does seem to be a politically tantalizing option for family equality in the near
future.
i.
A New Inclusive Vision of Social Security Family Benefits
Proposals to impose earnings sharing or privatize the Social Security system sit at two
extremes: on the one hand, the benefits of earnings sharing are confined to families formed by
marriage; on the other hand, a radical version of privatization would obliterate protection for any
families. A better, yet idealistic, solution might lie somewhere between these two poles. How
can the Social Security system achieve its stated purpose of “affording more adequate protection
to the family as a unit?”205
Could the system actually start moving in the direction of
recognizing more nonmarital relationships, as it has for illegitimate children? One way to
explore the possibilities for such reform is to compare Social Security with other public
203
Kijakaze, supra note 2 at 229-30. See also notes 67 and 106 and accompanying text for reasons why low-wage
workers receive sub-par Social Security coverage.
204
Given trends in federal law and the nationwide drive to preserve traditional marriage, it is unlikely that the
discrimination against same-sex marriages will end anytime soon. One major obstacle is the Defense of Marriage
Act (DOMA), which defines marriage as between a man and a woman, excluding same-sex couples from the
definition. DOMA also permits states to refuse to recognize marriages from other states so that no state would be
forced to extend rights to same-sex couples married in another state. Because DOMA is a federal law, even if a state
sanctions same-sex marriage, federal benefits, like Social Security, will not extend to cover same-sex couples. See
Liz Seaton, The Debate Over the Denial of Marriage Rights and Benefits to Same-Sex Couples and Their Children,
4 MARGINS 127, 144-46 (2004); Defense of Marriage Act, Pub. L. No. 104-199, 110 Stat. 2419 (1996); 28 U.S.C.
§1738C (2005)(providing that no state… is required to give any effect to a law that treats a “relationship between
persons of the same sex…as marriage”); 1 U.S.C. §7 (2005)(defining marriage as between a man and a woman).
205
Weinberger, 420 U.S. at 643.
50
assistance programs. Examining programs like welfare, food stamps, and public housing will
illuminate potential avenues for creating a more inclusive program that would better support all
families.
Comparing Social Security to Other Public Programs
Broadly speaking, public programs in the United States provide two types of assistance: cash
aid and non-cash goods, such as food or housing. While cash aid programs are generally quite
stringent in their administration and guidelines for eligibility, goods programs are more flexible.
As a result, programs that distribute goods typically adopt broader conceptions of the family than
programs that pay cash.
For example, the federal Food Stamps Program distributes benefits on the basis of a
“household” relationship, which is defined by people who live together and purchase and prepare
food together.206 In place of requiring state-sanctioned marriage or blood relationships, the Food
Stamps program will provide cash earmarked only for particular types of food purchases to
needy households headed by married couples, same-sex partners, cohabitants, and unwed parents
alike.
This functional approach, which is also used by public housing programs, is more
accepting of untraditional families than Social Security. The Food Stamps program also has
unlimited scope as an uncapped program, which might explain part of its liberal approach to
disbursing benefits.
On the other hand, Food Stamps only disburses benefits to needy
households. Perhaps because the families Food Stamps supports are so poor, the legislators and
policymakers who created the program begrudgingly accepted the fact that all qualified
recipients truly needed the food. Additionally, Food Stamps benefits are circumscribed in that
206
See 7 U.S.C.§2012 (2005). Note that Food Stamps and the Federal Supplemental Nutrition Program for Women,
Infants, and Children (WIC) still make use of family presumptions in eligibility determinations.
51
the only thing a recipient can purchase with their Food Stamps is food.207 Since the benefits are
so restricted, there may be fewer perceived problems with giving poor (read: irresponsible)
people food than giving poor people cash. In other words, policymakers can afford to be less
moralistic about defining eligibility for food programs because it is easier to dole out food than
cash, which is more likely to be spent in “unacceptable” ways.
In contrast to goods programs, cash benefits (e.g., Social Security and welfare) tend to define
“the family” in the most traditional ways. To qualify for Social Security children’s benefits, a
relationship between an insured parent and child must be created biologically (natural children),
by marriage (stepchildren), or by an approved adoption. For spousal benefits of any kind, the
relationship to the insured spouse must be founded on marriage. Generally, to qualify for
welfare there must also be evidence of a parent-child relationship (biological or adopted). While
welfare programs ordinarily cover unmarried parents and their children, “the family” and even
marriage still plays a key part in determining eligibility. Welfare recipients are sometimes
penalized for having additional children outside of marriage and may be required to participate in
programs that encourage marriage.208
Although programs that distribute cash tend to define the family in more traditional ways
than goods programs, recipients of Social Security benefits or welfare may spend their benefits
in whatever way they choose. Despite the flexible nature of cash benefits, welfare and the family
Social Security benefits created in 1939 have always been two wildly different systems of public
assistance for single mothers. Social Security benefits allowed qualified widowed mothers with
207
Food Stamps recipients cannot use food stamps benefits to purchase prepared foods or even essential household
items, like soap or toilet paper.
208
See e.g., the fourth purpose of the Personal Responsibility, Work, and Family Promotion Act of 2003, H.R. 4,
108th Cong. (1st Sess. 2003); Christine Heath, Marriage Promotion Debate Tied to Welfare, Washington Times,
June 28, 2004 available at http://washingtontimes.com/upi-breaking/20040628-04390-9224r.htm (last visited
January 1, 2006); Katherine Boo, The Marriage Cure: Is Wedlock Really a Way Out of Poverty? THE NEW
YORKER, August 18, 2003 at 105.
52
children to rely on an often-sizeable benefit while caring for children under the age of 18
regardless of their financial condition. Single mothers, who were divorced, deserted, or had their
children outside of marriage were left to rely on the moralizing welfare system. In stark contrast
to widows on Social Security, welfare recipients had to (and still must) meet stringent eligibility
standards and prove neediness.
Since Social Security’s inception the differences between welfare and Social Security have
become even clearer. As Social Security’s vision of the family gradually became more liberal
through judicial and legislative change, in recent years welfare has grown much more
conservative. In 1996, welfare reform replaced Aid to Families with Dependent Children
(AFDC) with state control to disburse federal Temporary Assistance to Needy Families (TANF)
grants. The move from AFDC to TANF established lifetime limits on receipt of benefits, gave
states the option to impose family maximums on families receiving welfare benefits, and
required that all recipients begin work or job training after two years on aid. 209 Mothers on
TANF are now almost universally required to work despite having young children in their care.
Women on welfare also face a five-year lifetime limit on their receipt of aid. Even child support
benefits are limited to many welfare families who often see only a $50 “pass-through” of the
support a father pays. Frequently, the rest of the child support payment goes to the government
to defray the cost of the welfare grant.210 By contrast, single mothers who receive Social
Security, whether widowed or divorced, can draw a parent’s benefit for as many as 16 years and
209
Rebecca M. Blank, Policy Watch: The 1996 Welfare Reform,
http://www.northwestern.edu/ipr/publications/nupr/nuprv02n1/blank2.html (last visited Nov. 12, 2005).
210
National Center for Children in Poverty, State Policy Choices: Child Support,
http://www.nccp.org/pub_csi04.html (last visited Nov. 12, 2005)(noting that many families receiving TANF
“benefit little, if any, from the child support collected on their behalf because states may opt to retain the money as
reimbursement for TANF benefits”).
53
benefits for their children for even longer.211 Mothers receiving Social Security are not required
to work and when they do work, they face better incentives than women receiving TANF. 212
In essence, by requiring TANF recipients to work but permitting caretaking parents who
collect Social Security to stay out of the workforce, welfare reform has partially inverted the
welfare system. While the expectation that “most women would participate in the labor force
gradually came to apply to women on welfare,” somehow parents receiving caretaker’s Social
Security benefits escaped this mandate.213 If welfare mothers were once considered lazy and
irresponsible, strict new welfare-to-work requirements have ended the illusion that receiving
welfare is easy or that it requires no work. Why hasn’t the stigma attached to receiving welfare
begun to dissipate in light of this fact? Is it sympathy for a widowed spouse’s loss that insulates
Social Security recipients from the moralizing critiques directed at single mothers on welfare?
Might the fiction that the deceased spouse has purchased “life insurance” by contributing to
Social Security form the basis for this disparate treatment? Perhaps the basis for this distinction
is as simple as the fact that a widowed wife relied on her husband while a single mother “chose”
to have a child without other support. Even if the rationale for maintaining this dual-system is
more inoffensive, welfare reform has made Social Security an even more preferable benefit in
comparison to other public assistance programs.
211
Also note that wives of insured disabled or retired workers are given the same preferential options to collect
caretaker’s benefits as widowed recipients. See Sugarman, Reforming Welfare, supra note 62 at 819.
212
For people receiving Social Security caretakers’ benefits, the impact earnings have on benefits varies based on
the amount earned (i.e., low-wage workers can retain more of their Social Security benefit than beneficiaries who
are earning more substantial wages). How Work Affects Your Benefits, SSA Publication No. 05-10069 at *4
available at http://www.ssa.gov/pubs/10069.pdf (last visited Nov. 12, 2005). In general, for parents under
retirement age, $1 in benefits is deducted for every $2 in earnings above an annual limit of $12,000 (in 2005). What
You Need to Know When you Get Retirement or Survivors Benefits, SSA Publication No. 05-10007 at *13 available
at http://www.ssa.gov/pubs/10077.pdf (last visited Nov. 12, 2005). While income disregards have improved after the
switch from AFDC to TANF for women on welfare, even under California’s liberal standards only the first $225 of
earned income is fully disregarded with 50 cents for every dollar earned above that threshold level counting against
the grant. CAL. WEL. & INST. CODE §11451.5 (2005).
213
BERKOWITZ, supra note 8 at 9-10
54
With these other programs in mind, it appears that Social Security, without any means
test, sits at the top of a benefits hierarchy since the program generally pays the highest benefits to
the people who may need the benefits the least. While Food Stamps and public housing may be
more accessible to nontraditional families than Social Security family benefits, the fact that
benefits are so meager and are available only to the very low-income make Food Stamps and
public housing less desirable. Welfare (TANF) might be at the very bottom of the hierarchy.
Although TANF resembles Social Security by distributing cash benefits, the program is available
only to the poorest families.
It also subjects recipients to demanding restrictions that are
completely absent for Social Security recipients.
How can Social Security reach more
nontraditional families (like public housing and food stamps) without losing its relatively revered
status as an earned-benefit? Is there a way for Social Security to widen its eligibility standards
by broadening its definition of the family without inheriting the stigma that is currently reserved
for other programs? Out of fairness, Social Security family benefits should extend to cover all
families that have a wage earner paying Social Security taxes. Any change to Social Security
should retain the feature of the program that indexes benefits according to income. This feature
makes Social Security an altogether different kind of program, insulates it from “welfare reform”
type attacks, and also permits the program to provide income to middle and low-income
individuals who might not qualify for welfare but still need Social Security benefits. In a radical
vein, Social Security could scrap defining “the family” along marital lines and adopt eligibility
standards founded on the notion of a “household,” like Food Stamps. A more practically feasible
proposal might take a two-step approach to insuring currently uncovered children of nonmarital
relationships and extending caregivers’ benefits to the “informal” (non-adoptive and nonbiological) parents of these children. Each of these options will be considered in turn.214
214
Thanks to Professor Sugarman for providing this interesting suggestion. I have adapted his idea somewhat but it
55
Thinking Outside the Marriage Box: Insuring “Households”
Converting Social Security from a family-based to a household-based program would
increase the flexibility of the program and would reap significant advantages for untraditional
families. The household model, for example, would cover children of unmarried partners who
had not been adopted by their non-biological parent if that parent retired or experienced death or
disability. All unmarried partners would also be able to provide their partners with retirement
and survivors’ benefits.215 Other dependents in the household, say a niece or nephew, or a
disabled sibling, might also qualify for benefits under the household scheme. Since Social
Security is indexed on an individual’s labor input, family benefits, even if based on a householdnotion would have to retain maximum family grants. This might help prevent formation of large
“households” for economic profit. The household model could also adopt earnings sharing so
that wage earners in a household would share equally in what was earned during the household
relationship and all earnings and benefits would be fully portable if a household relationship
dissolved.
While the changes that would increase coverage for family members seem positive, problems
might arise in defining the administrative scope of the “household” model. For example, the
household model would have to reckon with how to draw the line between family-like
relationships and more informal relationships, such as roommate living situations. Should all
members of a household automatically be deemed mutually dependent upon anyone living in the
house who also earns income? Such a rule would undoubtedly go too far since it would include
people who live together but in no way depend upon the other person for income or support. The
still forms the basis for this proposal.
215
Note that the limit in Boles that refused to extend benefits to an unmarried mother of an illegitimate child who
was not living with the insured parent would persist. What to do with the claim of a child, like Norman Boles, who
was also not living with his insured parent, would be a closer question. This might be resolved in favor of eligibility
despite the lack of a shared household because of the presumed obligation to pay child support.
56
line here could be difficult to draw. College roommates, who may live together for only a year
and earn spending money from part-time jobs might “share a household” but are in no way
financially dependent on one another. Two older single women who have lived together for a
decade, own a home together, and fully support each other would present a more compelling
case for inclusion in the “household” definition despite their lack of intimacy. Perhaps marriage
or domestic partnership, or caring for a child together, could create a presumption of dependency
that would qualify any spouse or partner for spousal benefits and any children in the home for
child’s benefits. Unmarried partners without children could be required to prove dependency in
the same way that stepchildren are required to demonstrate dependency. This solution would
still incentivize marriage and might appeal to people who want to retain preferences for
marriage.
The household model would encounter another significant problem when dealing with
dissolution of marriages and more informal relationships. As a legal contract, marriage has an
easily defined start and ending. Households, by contrast, are more informal. Thus, defining their
dissolution could prove administratively taxing. Finally, a “household” vision of Social Security
could open benefits up to polygamous households, which could also incite political opposition.
This abbreviated discussion of transforming Social Security from a family to a household
benefit has highlighted a few of the difficulties that would arise if it were enacted. Given the
scope of the problems and the potential that such a proposal would increase the cost of
administering the program, the “household” proposal is unfortunately far-fetched. However, a
move in this direction would take great strides in equalizing access to Social Security for all of
today’s varied family forms.216
57
A Two-Step Strategy to Insuring Nontraditional Families
Changing Social Security for the better does not require shaking it to its foundations,
however.
Relatively minor legislative changes could bringparents and children who are
currently uninsured under the Social Security umbrella. The following proposal would extend
benefits to children who are raised by a non-biological, non-adoptive, and non-step- parent but
who depend on their “informal” parent for support. The proposal would also provide these
informal parents with caretaking benefits if their partner (the child’s parent) were to die. This
proposal would stop short of extending spousal or survivor’s benefits to unmarried partners in
hopes that same-sex marriage would one day be legalized to provide the same effect and because
earning gaps between secondary and primary earners are likely to decline in future years.
The Court has consistently framed the purpose of the Social Security Act as a state-sponsored
method of family protection. When the Court has refused to extend benefits, it has typically
grounded their decision in the fact that “Congress sought to limit the category of beneficiaries to
those who actually suffer economic dislocation upon the death of a wage earner.”217 While the
Court upholds the marriage-based aspects of Social Security legislation on rational basis review,
the current system of family benefits is both over-inclusive and underinclusive. As earlier
sections of this paper have demonstrated, by often using marriage as a proxy for eligibility and
dependency, the system sometimes overcompensates certain families at the expense of others
and even pays out money to people who could get by without it. Despite its unfairness, this
over-inclusiveness results from the “earned” aspect of the program, which is probably
responsible for Social Security’s enduring success.
As an earned benefit, Social Security
insurance is insulated from the stigma that is directed at means-tested welfare programs. Even so,
the program is also underinclusive because it does not cover everyone who might be dependent
217
Boles, 443 U.S. at 289.
58
on an insured wage earner. Notably, unmarried partners, some children of unmarried partners,
and other potential dependents are left out of family insurance coverage. With these uncovered
individuals in mind, it is clear that Social Security is not living up to its stated goal of “protecting
the family unit.”
To protect children living in families not founded on marriage, the Social Security program
could adopt the stepchild eligibility test for children who hope to receive benefits from informal
parents. This initial change would require such children to prove that they were “living with” the
informal parent and “dependent” upon the informal parent for support. Such a change would
create a class of “de facto stepchildren” who although never formally adopted, were nevertheless
parented and supported by their informal parent (the primary beneficiary). Including these
children would be politically attractive for a couple of reasons. Chiefly, this would probably
extend to relatively few children and, as a result, would probably result in little extra cost or
administrative burden. After all, the change would merely apply a test that the Social Security
Administration is already familiar with using for stepchildren. Additionally, this change would
provide income to children who would benefit from the money and who cannot be blamed for
the decisions of their parents. This “innocence” rationale might make such a proposal attractive
to policymakers and legislators.
Next, a statute could be adopted to award parenting benefits to the informal or “de facto”
parents who cared for their non-adopted, non-biological, and non-step child if the biological or
adopted parent of the child passed away. Such a change would benefit same-sex couples who
have been unable to adopt children that one partner brought in from a prior relationship. It
would also benefit blended families where non-biological parents are prevented from adopting
the children of their partner or spouse. This statute could also be used to award parenting
59
benefits to parents of illegitimate children who qualified for child’s benefits when, for example,
their biological parent died. Such a statute would overturn Boles, might greatly improve the
economic status of many single parents (primarily mothers), and could move families off of
welfare.
Finally, this proposal would not extend benefits to unmarried partners. A statutory proposal
to provide same-sex couples and cohabitants with equal access is likely to meet with major
opposition given the current hostility to recognizing same-sex relationships on a federal level.
Since spousal benefits will be extended to same-sex relationships if same-sex marriage is
legalized, it would probably be prudent to wait for that development instead of pushing for
unpopular legislative change now. Additionally, if wage disparities between men and women
shrink, fewer married women will qualify for spousal benefits in the future. It would make sense
to refrain from enacting more marital privileges even as the marriage benefit is waning for many
modern couples.218 All these changes would stop short of extending benefits to affinal relatives
(like siblings who receive care from adult siblings and would receive coverage under the
household proposal) or friends, but would take important steps to insuring families that are
currently given short shrift from the marriage-focused Social Security system. The legislative
changes would probably be upheld in Court since they would reflect Congress’ rational choice to
“concentrate limited funds where the need is likely to be greatest.”219 By applying a dependency
test to “de facto” stepchildren the new changes would be especially apt to award benefits only to
those who needed them. Finally, the relatively limited scope of these changes might make them
appealing to legislators who want to avoid incurring greater Social Security costs.
218
Statistics on earnings discrepancies in same-sex couples were unavailable. However, assuming the wage gap is
gendered or results from traditional gender norms, same-sex couples are also less likely to consist of primary and
secondary earning partners.
60
Conclusion
In several ways, Social Security supports a type of family that is increasingly a relic. It
fails to accommodate a growing number of American families by basing family benefits solely
on marital or biological connections and by rewarding traditional family structures with
proportionately greater benefits than families that break up or break from the two-parent, single
earner mold. Perhaps Social Security would be better off privatizing, thus allowing individuals
to choose who they will support free from government control and the presumption that only
marriage will suffice for derivative benefit eligibility. The appeal of privatization, however, is
undercut by the problems it would pose for some of the most vulnerable families in our society.
In light of these concerns, this paper has offered an ambitious suggestion for reform,
which would adopt the “household” as the relevant eligibility marker for family benefits instead
of marriage, adoption, or a biological connection. This proposal would take a giant leap in terms
of covering all families but is undercut by familiar concerns about administration and cost. A
more moderate proposal, to insure parents and children that currently fall into a gap in Social
Security family benefits has some possibility for success because it imports elements of the
program that are already in use in other areas. Despite these prospects for reform, Social
Security is a large and entrenched program and ideas for reform are often met with opposition.
After all, Social Security is a unique public program. Unlike other forms of government aid,
Social Security pays qualified workers benefits based on money they have already contributed.
The final proposal forwarded in this paper would offer family benefits on an equal basis to
people who currently pay Social Security taxes and whose families should receive their fair
reward. Unfortunately, any step towards including more workers and dependents is especially
likely to get shut down given concerns not just about maintaining traditional values but because
61
of ongoing worries about the financial soundness of the program itself. To make this proposal
more viable, additional research defining the costs that would be incurred from such a change
and the benefits that would adhere to formerly uninsured families will be necessary.
62