Social Security Benefits for Spouses*

Peter W. Martin

[Warning: I am working on updating this ancient piece. At the moment, the text is far ahead of, and in some cases out of step with, the footnotes in this revision process.]

The fundamental problem with the current social security system is its inability to cope with the employment patterns of the majority of women, who are neither full-time homemakers or full-time employees.

Staff memorandum, Justice Department Task Force on Sex Discrimination.⁽¹⁾

Social Security is not something that is dealt with once and for all by just one bill. We will be voting on social security bills around here as long as we are Members of this Body, I would think.

Senator Russell B. Long on the floor of the Senate, December 15, 1977.⁽²⁾

Table of Contents

Introduction

I. The Origin of Spouse Benefits

II. The Failure of Retired-Worker Benefits to Displace Spouse Benefits

III. Spouse Benefits for Men and Constitutional Equality

IV. The Congressional Response to Goldfarb

V. The Essential Arbitrariness of Spouse Benefits

A. Should All Who Are Legal Married at Benefit Time Receive Spouse Benefits?

B. Should Others--De Facto Spouses and Former Legal Spouses--Receive Spouse Benefits?

C. Should More Than One Spouse Receive Benefits on a Single Worker's Account?

D. The Remarriage Dilemma 824

VI. Improving Spouse benefits as a Technique for Improving Benefits for Women

VII. An Alternative: Recognition of Marriage on the Contribution Side

A. Retired-Worker Benefits and Women

B. The Community-Property Model

C. Earnings-Splitting Applied to Social Security

Conclusion

Introduction

A few basic facts about the OASDI structure demonstrate why true Social Security benefit equality between men and women is so difficult to achieve and how spouse benefits relate to that problem. Two separate routes to entitlement exist: through one's own prior covered employment and through the prior covered work of a family member, most commonly a spouse or parent. Primary benefits, based on one's own work, become available upon retirement (as measured by advanced age and, for a slightly younger group, age and low earnings), or earlier in the event of total disability. Secondary benefits, arising from another's work, are paid to "dependents" or "dependent survivors" of covered workers. Those eligible include several categories of both children and adults. Among adult dependents, spouses are by far the most numerous.

The Social Security Act currently provides secondary benefits to the wives or widows of covered workers who retire, become disabled, or die. To qualify, a woman must have been married to the worker for a minimum period and must be old (with a disabled widow the threshold drops to 50) or caring for children under sixteen. If a wife's or widow's primary retired-worker or disability benefits equal or exceed her secondary benefit entitlement, she receives only the primary benefits. However, if her secondary benefit amount is greater she receives both her primary benefit and enough of the secondary benefit to bring the total up to its larger figure.

Men can also qualify for benefits based solely on their status as husband or widower of a worker; but spouse benefits go overwhelmingly to women. Table 1 indicates the relative importance of primary and secondary benefits to men and women. Nearly all adult male beneficiaries receive primary benefits, but almost half of adult female beneficiaries receive secondary benefits.⁽³⁾ Average monthly payments to male retired workers are substantially higher than those to female retired workers, spouses of male retired workers, or widows. Because of this disparity more than 40% of the women receiving retired worker benefits also receive secondary benefits. The number of men with "dual entitlement" is miniscule, less than 1%.⁽⁴⁾

Table 1

Adult Social Security Beneficiaries by Sex and Benefit Type

December 2002

	Men		Women
	N	$	N	$
Primary:
Retired Workers	15,107,850	1,008.10	14,095,810	773.90
Disabled Workers	3,069,140	935.00	2,466,720	708.60
Total	18,176,990		16,562,530
Secondary:
Spouses of Retired Workers
--without children^a	34,290^c	257.30	2,593,750	456.30
--with children^b	----	----	53,050	360.90
Spouses of Disabled Workers
--without children^a	4,110^c	168.10	49,060	286.50
--with children^b	----	----	98,090	175.90
Surviving Spouses
--without children^a (nondisabled)	39,800	668.00	4,523,860	863.10
--without children^a (disabled)	5,820	380.10	200,050	551.10
--with children^b	11,130	546.30	182,140	642.20
Total	95,150		7,700,000

N = Number receiving benefits.
$ = Average monthly benefit.

Source: Social Security Administration, Annual Statistical Supplement, 2003, Tables 5.A1.1, 5.A1.3, 5.A1.5.

^a Spouses eligible by virtue of age.
^b Younger spouses eligible because of eligible children.
^c The figure represents all husbands receiving this type of secondary benefit, including those eligible because of children.

Social Security benefits of both types, primary and secondary, hold greater importance for women than men. As the Social Security Administration points out:

• For unmarried women – including widows – age 65 and older, Social Security comprises 52 percent of their total income. In contrast, Social Security benefits comprise only 38 percent of unmarried elderly men's retirement income and only 35 percent of elderly couples' income.
• Social Security is the only source of retirement income for 29 percent of unmarried elderly women.
• Elderly women are less likely than elderly men to have significant income from pensions other than Social Security. In 2002, only 21 percent of women aged 65 or older were receiving their own pensions (either as a retired worker or survivor), compared to 28 percent of men.⁽⁵⁾

It would be comforting to suppose that changes in the labor market and allocation of work within the family home will change this picture in time, but if so that will not be any time soon. Projecting into the foreseeable future, a 1998 interagency working group concluded:

• The percentage of women receiving benefits based solely on their own earnings history is expected to rise from 37 percent today to 60 percent in 2060. However, this means that 40 percent of women will continue to receive benefits based on their husband’s earnings.
• The average full monthly benefit for retired female workers based on their own earnings record, which is currently 62 percent of the average for men, will rise to 67 percent in 2050.
• The difference in life expectancy at age 65 between men and women will fall only slightly under Social Security Administration projections from a gap of 3.6 years today to 3.4 years in 2030. Thus, in the future, women will continue to depend on Social Security for more years than men will.⁽⁶⁾

In short, any effort to reform Social Security must take account of the importance of the program to woman, the role of the program's benefits based on a past or present marriage in reducing the benefit disparity between men and women, and the implausibility of expecting that social or economic changes alone will solve this issue of gender equity.

This essay focuses on the law that governs entitlement to and the amount of spouse benefits, tracing the present statutory provisions from their origin through 2004. It considers why spouse benefits continue to have such major importance for women despite substantial gains in female participation in the labor force. Although this study does not fully explore alternative means of improving Social Security benefits for women nor the implications for women of Social Security restructuring proposals of other sorts, it does attempt to illuminate those larger issue by exploring the inherent limitations of dependent benefits. Comparison with a differing approach to marriage and benefits, modeled on state community-property and divorce law, highlights those limitations.

I. The Origin of Spouse Benefits

Originally, Social Security did not include spouse benefits. The Social Security Act of 1935 tied benefit entitlement rather closely to contribution; the program paid benefits only to retired workers (male or female) or, upon death, to the worker's estate.⁽⁷⁾ In 1939, however, acting on the recommendation of the Social Security Board,⁽⁸⁾ Congress relaxed the benefit-contribution relationship in several ways, including the addition of secondary benefits.⁽⁹⁾ The amendments of that year established "supplemental payments" to aged wives of retired workers to "take account of [the] greater presumptive need of the married couple without requiring investigation of individual need."⁽¹⁰⁾ On similar "welfare" grounds, the amendments substituted survivor benefits to aged widows and children for lump-sum payments to the worker's estate.⁽¹¹⁾

An acute short-term economic and political problem – the extremely low payout and coverage of Social Security during its infancy – motivated enactment of these measures; they were thought to possess limited long-range importance: "Since in the course of time many women will have developed substantial benefit rights based upon their own past earnings, the cost of providing the supplement for dependent wives [and widows should] gradually decline, and eventually the additional cost [should] be reduced to a relatively small amount."⁽¹²⁾ The inequities thus created – single workers and two-worker couples contributed no less and sometimes more than one-worker couples, yet reaped no advantage from the new classes of benefits – seemed a small matter at a time when the Social Security tax was only one percent of the first $3,000 of annual wages.⁽¹³⁾

Anticipating objections to secondary benefits on the ground that "presumptive need" should not shape entitlement, the Social Security Board argued that the program already reflected a concern for relative need.⁽¹⁴⁾ The Board suggested that providing new dependent and survivor benefits and revising the retired-worker benefit formula did not violate the contributory or "insurance" character of the system so long as a "reasonable relationship . . . between benefits payable and past earnings."⁽¹⁵⁾ was maintained. The Board found no offense to this "reasonable relationship" test in the additional wife benefits because everyone, married or single, would, for the immediate future, receive more than they had contributed.⁽¹⁶⁾ The 1939 amendments also set a maximum on the total monthly benefits payable on a particular worker's account. That ceiling, today termed the "family maximum," limited the degree of unfairness to individual contributors resulting from the new secondary benefits.⁽¹⁷⁾

In light of more recent current controversies over Social Security financing it is worth noting how Congress planned to finance the increments of greater "adequacy" added in 1939. Initially, the large imbalance between taxes and benefits provided a ready source of funds; the amendments were in large part a response to that surplus. But what of the future, when a significant percentage of the retired population would draw benefits? The Board's 1939 report dealt candidly with this issue:

The tax provisions embodied in the present law would probably cover the increased annual cost for the first 15 years. They would also probably provide a small reserve, which would be invested and earn some interest. But when future annual benefit disbursements exceeded annual tax collections plus interest earnings, some other provision would have to be made for the funds which, under the existing plan [established by the 1935 Act], would be secured from interest on accumulated reserves. It would then be necessary to do one of two things: increase the pay-roll tax, or provide for the deficiency out of other general taxes.⁽¹⁸⁾

The original terms governing wife and widow benefits left little doubt that Congress intended them to meet a presumed loss of income caused by the retirement or death of a husband rather than has an expression of the notion that wife had a stake in her husband's earnings. Eligibility hinged on marriage at the time of the worker's retirement or death, not during the period in which he built up his Social Security wage account.⁽¹⁹⁾ To qualify in accordance with the presumptive need rationale, a woman had to show she was living with her husband or at least financially dependent on him.⁽²⁰⁾ Regardless of her age, a wife could receive spouse benefits only when her husband was old enough to receive retired-worker benefits and had, in fact, retired and applied.⁽²¹⁾ Divorce ended a wife's entitlement; remarriage ended a widow's.⁽²²⁾ And each dollar of retired-worker benefits a woman was eligible to receive on her own account displaced a dollar of these need-premised spouse benefits.⁽²³⁾

On the other hand, it then seemed obvious that most eligible wives and widows would have been married to the covered worker during the period in which he earned his primary benefits.⁽²⁴⁾ The new secondary benefits reflect, ever so slightly, a notion of shared contribution. Rather than adding a single uniform payment for dependent wives and surviving widows (which would have typified a need-oriented or "welfare" approach), the benefit formula, set individual amounts proportionate to the retired-worker benefits a woman's husband could claim at sixty-five. A dependent wife was entitled to fifty percent of her husband's benefits; a widow, seventy-five percent.⁽²⁵⁾ Despite frequent references then, as now, to "the couple's benefit" or to "a supplement for married men," the secondary benefits established in 1939 truly belonged to the eligible spouse. The male retired worker received his benefit; his eligible wife received her own.⁽²⁶⁾ However, the number of points at which the wife's eligibility rested on the covered worker's status – including matters largely within his control – significantly qualified this independence.

During their first decade, wife and widow benefits proved to be expedient need-focused or antipoverty measures. Although the male-female ratio was more closely balanced than at present, even during the forties elderly women outnumbered elderly men.⁽²⁷⁾ Very few older women had enough recent employment (post-1936) to qualify for Social Security retired-worker benefits.⁽²⁸⁾ Poverty was so prevalent among the elderly that any program paying benefits to those 65 and over could reasonably have been called an antipoverty measure.⁽²⁹⁾ Poverty hit hardest among elderly women,⁽³⁰⁾ particularly widows.⁽³¹⁾

Under these circumstances, paying secondary benefits to wives and widows of Social Security contributors seemed a sensible way to provide limited amounts of monthly income to a class of needy people. In 1942, for example, only 13,000 women qualified for retired-worker benefits, compared to 87,000 men.⁽³²⁾ Wife benefits permitted an additional 33,000 elderly women to qualify for payments; 15,000 elderly widows collected on the accounts of their deceased husbands, and 32,000 widows under 65 were eligible because of young children of deceased workers.⁽³³⁾ Because payments were small,⁽³⁴⁾ the few cases where presumed need did not in fact exist represented neither serious fiscal waste to the system nor large windfall to the recipient.

Even with the new benefits, Social Security reached less than twenty percent of the nation's elderly by the end of the forties (see Table 2). Without benefits for dependent and surviving spouses, the program's payments in 1949 would have diminished by nearly one quarter and the number of beneficiaries by an even larger fraction.⁽³⁵⁾

Table 2

Aged Social Security Beneficiaries
by Sex and Benefit Type, 1950-2000

	Population^a,b		All Soc. Sec. Beneficiaries^b,c	Retired-Worker Benefits^a,b
	Men	Women		Men		Women
			%	%	N	%	N
1950	5,857	6,541	16	25	(1,469)	5	(302)
1960	7,542	9,133	62	69	(5,217)	27	(2,487)
1970	8,405	11,680	86	85	(7,111)	43	(5,010)
2000	14,400	20,600	93	93	(13,442)	61	(12,511)

	Average Monthly Retired-Worker Benefits			Spouse Benefits^a,d		Surviving-Spouse^a,e Benefits
	Men	Women	Ratio Women/Men	Men	Women	Men	Women
1950	$45.67	$35.05	.77	1	499	--	314
1960	$81.87	$59.67	.73	15	2,144	2	1,542
1970	$130.53	$101.22	.78	8	2,492	3	3,175
2000	$950.00	$706.00	.74	34	2,773	40	4,770

Sources: Bureau of the Census, U.S. Dep't of Commerce, Social and Economic Characteristics of the Older Population, 1974, Table 1, at 3 (Current Population Reports, Series P-23, No. 57, 1975); Lisa Hertzel & Annetta Smith, The 65 Years and Over Population: 2000, 3 (Bureau of the Census 2001); Soc. Sec. Bull.--Annual Statistical Supplement, 1975, Table 23, at 57, Table 84, at 121, Table 85, at 122, Table 100, at 131, Table 107, at 136; Social Security Administration, Annual Statistical Supplement, 2003, Table 5.A10.

^a Numbers in thousands.
^b Figures represent those 65 years old and older.
^c Figures represent those receiving either primary or secondary benefits.
^d Figures represent those 62 years old and older receiving benefits based on the account of a retired worker.
^e Figures represent those 60 years old and older.

II. The Failure of Retired-Worker Benefits To Displace Spouse Benefits

Contrary to congressional expectations on enactment, spouse benefits have remained significant even as more women have become eligible for retired-worker benefits on their own account. Labor force participation by women has risen steadily since 1939,⁽³⁶⁾ yet spouse benefits have continued to grow in recipient number and total dollar amount (see Table 2).⁽³⁷⁾ Many of the reasons for this paradoxical phenomenon could not have been foreseen in 1939. Under the 1939 formula, a woman retiring at age 65 in 1949 had only to have earned average monthly wages of $37.00 for 12 years to receive benefits on her own account, so long as she was married to the average male beneficiary retiring the same year.⁽³⁸⁾ The average monthly wages of a female worker in manufacturing in 1939 were almost twice that amount.⁽³⁹⁾ The gradual disappearance of spouse benefits was thus a reasonable prediction assuming stable wage levels and no revisions of the benefit formula.

Events soon upset these underlying assumptions. From 1940 on, wages and prices rose steadily and, beginning in 1950, Congress responded with periodic revisions of the benefit formula.⁽⁴⁰⁾ Each revision increased retired-worker benefits and produced proportionate increases for secondary-benefit recipients. The relationship that evolved between the two types of benefits differed substantially from that contained in the 1939 legislation. In addition, the same motivations that led to the creation of dependent and survivor benefits later induced Congress to liberalize the terms on which secondary benefits were awarded. Congress passed at least ten such amendments between 1950 and 1977.

In 1950, for example, Congress granted wife benefits to younger wives caring for eligible children of retired workers, and extended survivor benefits to divorced former wives caring for children of deceased workers.⁽⁴¹⁾ A 1956 amendment reduced from sixty-five to sixty-two the age at which women could collect wife or widow benefits (without having young children in their care), or retired-worker benefits.⁽⁴²⁾ For widows, a 1965 amendment reduced the age to sixty;⁽⁴³⁾ two years later it was dropped to fifty for totally disabled widows.⁽⁴⁴⁾

Congress dropped the "living with" requirement for both wives and widows in 1957, leaving marital status the sole test of "dependency"; the marital status definition was also clarified slightly.⁽⁴⁵⁾ Three years later, Congress again liberalized the wife- and widow-benefit provisions. Legislation reduced the durational requirement a wife had to meet from three years to one.⁽⁴⁶⁾ Moreover, a new purely federal test of marital status qualified those who, though not legally married under state law, had gone through a marriage ceremony in good faith.⁽⁴⁷⁾

In 1958, two years after adding disability benefits to Social Security, Congress extended dependent benefits to spouses and children of disabled workers on the same terms as it had to dependents of retired workers.⁽⁴⁸⁾

The 1939 amendments pegged wife benefits at 50% of the worker's primary insurance amount and aged-widow benefits at 75%.⁽⁴⁹⁾ In 1961, the latter figure was increased to 82.5% and in 1972, to 100%.⁽⁵⁰⁾

Finally the 1965 amendments extended the equivalent of wife and widow benefits to divorced women whose marriages had lasted twenty years and who met a test of continuing actual dependency on their former spouses.⁽⁵¹⁾ Congress removed the actual-dependency requirement in 1972,⁽⁵²⁾ and reduced the durational requirement to ten years in 1977.⁽⁵³⁾

These numerous amendments, combined with the growth in Social Security coverage, made it increasingly easy for women to qualify for spouse benefits. The dramatic improvement in basic benefits and in the widow's entitlement as a percentage of the primary insurance amount substantially raised the value of such secondary benefits. Although more women qualify for retirement benefits, the present retired-worker formula, together with women's wage and work patterns, often keeps those benefits below wife or widow entitlements (see Figure 1). Over twenty percent of the women currently receiving retired-worker benefits also receive secondary spouse benefits, compared to fifteen percent at the end of 1966 and less than ten percent at the end of 1956 (see Figure 1).

Behind the many spouse-benefit amendments lay no clear scheme or consistent rationale. Indeed, Congress's incremental approach to modifying these provisions virtually guarantees that "[no] particular amendment fits with mathematical nicety into a carefully conceived overall plan for payment of benefits."⁽⁵⁴⁾

Figure 1

Source: Social Security Administration, Fast Facts and Figures About Social Security 22 (2003).

Nonetheless, taken together the amendments reflect two important features of the spouse-benefit approach. First, because entitlement to such benefits stems from a "presumed need" rather than a contribution rationale, Congress has frequently succumbed to the argument that a particular group of women (or occasionally men) excluded by existing eligibility rules should be included. The argument for their inclusion is simply that they are at least as deserving as those already receiving benefits. Because the line between eligibility and ineligibility is demonstrably arbitrary, it has proven, over time, to be particularly unstable.

Second, the post-1939 amendments seemed to respond to the entanglement of spouse benefits in hard-to-manage factual and legal questions. The dependent and dependent-survivor benefits added in 1939 had proved complicated to administer. Eligibility and benefit calculations for retired workers primarily required proof of age and use of routinely collected wage records. In contrast, to determine eligibility for wife benefits required ascertaining the validity of marriages and divorces under state law, and scrutinizing living and support arrangements. As a result, these secondary benefits accounted for a significant portion of Social Security administrative appeals and litigation⁽⁵⁵⁾ before Congress injected the more troublesome issue of "disability" into the system in 1956.⁽⁵⁶⁾ Removal of the "living with" test for wives and widows in 1957 and the support test for divorced wives in 1972 eliminated many difficult legal and factual issues.⁽⁵⁷⁾ The 1960 addition of a liberal federal definition of "spouse" also simplified marital status determination in some cases.⁽⁵⁸⁾

Any conclusion about what the many spouse-benefit amendments demonstrate is debatable; their cumulative effect, however, is clear: they assured that spouse benefits would not "wither away."

III. Spouse Benefits for Men and Constitutional Equality

Congress provided some secondary benefits to aged husbands and widowers as early as 1950.⁽⁵⁹⁾ They, however, carried a test of "actual dependency" on the wage earner more stringent than any applied to wives or widows.⁽⁶⁰⁾ That test remained after the 1957 amendments eliminated the requirement that wives or widows either be living with or financially supported by their husbands.⁽⁶¹⁾ It remained after 1972, when Congress deleted the requirement that divorced women be financially dependent on their retired or deceased former husbands.⁽⁶²⁾ Thus, by 1975, women received spouse benefits based solely on their marital status at benefit time; men did not. In addition, certain spouse benefits – principally benefits for divorced women and for young widows caring for children of deceased wage earners (mother benefits) – remained wholly unavailable to men.

In 1975, the Supreme Court began to chip away at this sex-differentiated structure in Weinberger v. Wiesenfeld,⁽⁶³⁾ holding that Social Security had to furnish young widowers with benefits comparable to mother benefits. Two years later, the Court substantially completed the job in Califano v. Goldfarb,⁽⁶⁴⁾ Califano v. Silbowitz,⁽⁶⁵⁾ and Califano v. Jablon,⁽⁶⁶⁾ ruling that Social Security must grant elderly widowers and husbands spouse benefits on the same terms it does to widows and wives.⁽⁶⁷⁾

In Wiesenfeld, the Court examined the benefits paid young widows (but not young widowers) caring for children from both the contribution and benefit side. On the contribution side, the distinction denied "women [workers the] protection for their families which men receive as a result of their employment."⁽⁶⁸⁾ On the benefit side, the Court rejected the view that benefits for young widows with children implemented a "policy of cushioning the financial impact of spousal loss upon the sex for which that loss imposes a disproportionately heavy burden."⁽⁶⁹⁾ It interpreted mother benefits as a measure focused on the welfare of children rather than widows. This led directly to a finding of constitutional invalidity: "Given the purpose of enabling the surviving parent to remain at home to care for child, the gender-based distinction of § 402 (g) is entirely irrational. The classification discriminates among surviving children solely on the basis of the sex of the surviving parent."⁽⁷⁰⁾

Goldfarb posed a harder case for two reasons. First, aged widowers could receive benefits (young widowers with children could not) if they could show actual dependency. Second, the lack of such a dependency test for widows could more easily be characterized as "an attempt to provide for the special problems of women."⁽⁷¹⁾ Nevertheless, by a 5 to 4 vote,⁽⁷²⁾ the Court held the differential treatment unconstitutional. The plurality opinion by Justice Brennan, joined by Justices White, Marshall, and Powell, focused first, like Wiesenfeld, on the contribution side. It suggested, however, that analysis need not go further, that a provision denying women wage earners the protection enjoyed by male contributors to the system was necessarily infirm.⁽⁷³⁾ Yet the plurality did look at the benefit side:

We conclude . . . that the differential treatment of nondependent widows and widowers results not, as [the Government] asserts, from a deliberate congressional intention to remedy the arguably greater needs of the former, but rather from an intention to aid the dependent spouses of deceased wage earners, coupled with a presumption that wives are usually dependent . . . .The only conceivable justification for writing the presumption of wives' dependency into the statute is the assumption, not verified by the Government . . . , but based simply on "archaic and overbroad" generalization . . ., that it would save the government time, money, and effort simply to pay benefits to all widows, rather than to require proof of dependency of both sexes. We [hold] . . . that such assumptions do not suffice to justify a gender-based discrimination in the distribution of employment-related benefits.⁽⁷⁴⁾

Justice Stevens supplied the necessary fifth vote. His opinion, like that of the dissenting Justices, focused on the benefit side since he found no merit in the argument that the statute discriminated against women contributors:

At the same salary level, all workers must pay the same tax, whether they are male or female, married or single, old or young, the head of a large family or a small one. The benefits which may ultimately become payable to them or to a wide variety of beneficiaries – including their families, their spouses, future spouses, and even their ex-wives – vary enormously, but such variations do not convert a uniform tax obligation into an unequal one.⁽⁷⁵⁾

Justice Stevens concluded that the preferential treatment of widows was neither the result of a considered effort by Congress to achieve administrative ease nor "the product of a conscious purpose to redress the 'legacy of economic discrimination' against females."⁽⁷⁶⁾ Instead, he found "that this discrimination [was] merely the accidental byproduct of a traditional way of thinking about females."⁽⁷⁷⁾ Conceding that the same reasoning applied to the statute upheld in Kahn v. Shevin,⁽⁷⁸⁾ Justice Stevens nevertheless considered the more recent and factually closer Wiesenfeld case controlling.⁽⁷⁹⁾

Three weeks later, in Silbowitz⁽⁸⁰⁾ and Jablon,⁽⁸¹⁾ the Court affirmed, without opinion, district court decisions that struck down on fifth amendment grounds the requirement of acutal dependency imposed on husbands but not wives. The Court simultaneously upheld, in Califano v. Webster,⁽⁸²⁾ the statutory formula that allowed women retiring before 1972 to omit three more low-wage years than men in calculating average monthly wages for retirement benefit purposes.⁽⁸³⁾ The Court characterized that formula as a permissible measure aimed at "redressing our society's longstanding disparate treatment for women."⁽⁸⁴⁾

IV. The Congressional Response to Goldfarb

By extending secondary benefits to men on the same terms as to women, Goldfarb and the other husband-benefit cases highlighted a growing problem with that scheme.⁽⁸⁵⁾ The attenuated chain of presumption it embodied – a presumption of need resting upon a presumption of financial dependency on the covered spouse's earnings which in turn rested upon marital status alone – too often failed to match the facts. The payment of benefits to those for whom the presumption of need and dependency was invalid could no longer be dismissed, as it was in 1939, as involving neither significant waste to the system nor major windfall to the recipient.⁽⁸⁶⁾

Two features of the spouse-benefit provisions helped to contain the problem. First, greatly expanded Social Security coverage made it more likely that a wife's work outside the home would be reflected in a Social Security wage record. Wives with sufficient wage records to support retired-worker benefits receive those benefits, which then offset spouse benefits dollar for dollar.⁽⁸⁷⁾ Second, the retirement test, applicable to both retired-worker and spouse benefits, blocked benefits to those who were non-needy because of continuing earnings of their own or their spouses, at least prior to age seventy-two.⁽⁸⁸⁾

Nonetheless, spouses who have never been financially dependent on a covered worker can and do receive substantial spouse benefits. That was true of some wives and widows before Goldfarb. But the husband- and widower-benefit cases threatened the system with a wave of new benefit claims, many of them from "nondependent" men who lacked Social Security retirement coverage of their own, not because of insubstantial employment, but because they had been covered by a public pension scheme other than Social Security. (Federal employees and many state and local government workers were at the time still uncovered by Social Security.)⁽⁸⁹⁾ It was more than coincidental that all three men seeking spouse benefits in the cases decided by the Supreme Court in March 1977 were retired federal employees on pension.⁽⁹⁰⁾

Responding to Goldfarb, the administration proposed a new test of actual dependency for all spouse benefits. That test would have qualified only people who earned less than their spouses during the three-year period immediately preceding the event triggering eligibility-retirement, death, or disability.⁽⁹¹⁾ Opponents argued that the proposal threatened to deprive many deserving women of benefits. They noted that a three-year period of relative earnings would often fail to reflect accurately the long-term economic relationship of a marriage, especially since the critical period under the test was likely to come between ages fifty-five and sixty-five when the probability of health or employment problems was high.⁽⁹²⁾

Congress substituted a narrower and less controversial provision. It offset pensions from uncovered public employment against secondary benefits in the same way the system has always offset retired-worker benefits. The provision, which did not extend to spouse benefits already applied for, created a reduction in spouse benefits equal to

the amount of any monthly periodic benefit payable to such [spouse] for such month which is based upon [his or] her earnings while in the service of the Federal government or any State (or political subdivision thereof. . .) if, on the last day [he or] she was employed by such entity, such service [was not covered by Social Security].⁽⁹³⁾

Congress was primarily concerned with preventing a post-Goldfarb flow of spouse benefits to "non-needy" male spouses. It was also hesitant to upset the expectations of women facing imminent retirement. For this reason, the amendment deferred the effective date of the provision for all women but for only a few men. The sex distinction was drawn indirectly; an exception to the new scheme applied to those who met the requirements of the spouse-benefit provisions as they were "in effect and being administered in January 1977" – in other words, prior to Goldfarb, Silbowitz, and Jablon.⁽⁹⁴⁾ No offset applies to such individuals if they became entitled to pensions from uncovered public employment during the five-year period following enactment.⁽⁹⁵⁾ Once the phase-in period passed, spouse benefits were limited to people without comparable publicly provided retirement benefits, whether in the form of retired-worker benefits from work covered by Social Security or pension benefits from public employment.

V. The Essential Arbitrariness of Spouse Benefits

Blocking the flow of spouse benefits to the largest visible group of "non-needy" recipients, although a constructive step, left untouched the principal source of arbitrariness in the spouse-benefit system. That arbitrariness lies in the provisions that establish which women (and men) have a sufficient connection to a covered former worker to qualify for spouse benefits, and in the related provisions that determine the amount of those benefits. Those provisions incorporate no coherent rationale for entitlement. A reasonable presumption of need, occasioned by the cessation of earnings on which the individual was dependent, no longer explains the statutory pattern. Many categories of spouses are eligible despite their lack of dependence on the insured's earnings at retirement, disability, or death. Similarly, constructive contribution for periods of work within the home, invoked from time to time to justify spouse benefits, finds no consistent expression in the eligibility rules.

Under any likely rationale, current spouse-benefit provisions distribute potentially large benefits to some who seem undeserving, while denying them to more appealing claimants. This flaw is not easily cured. The rules can be tinkered with; their history is one of constant tinkering. But so long as one set of principles determines contribution and entitlement to primary benefits, while a completely different set – focusing on family relationships at some later point or period – determines entitlement to spouse benefits, serious arbitrariness is inescapable. The follow eligibility issues, which have troubled the spouse-benefit system ever since 1939, stem ultimately from that fact.

A. Should All Who Are Legally Married at Benefit Time Receive Spouse Benefits?

The current legislation allows all those legally married to a covered worker at benefit time to receive spouse benefits, provided they meet a limited durational test. For wife (or husband) benefits, the marriage must have lasted at least one year at the time of application; for widow (or widower) benefits, at least nine months at the time of the worker's death.⁽⁹⁶⁾ The durational requirement has shrunk to its current insignificance over time.⁽⁹⁷⁾ Earlier requirements that the legal spouse be living with or financially dependent upon the covered worker have also vanished.⁽⁹⁸⁾

This heavy reliance on marital status has two consequences that pose serious questions of "fairness". First, Social Security provides spouse benefits to people who have neither been married to workers over any significant period of contribution nor built up any significant financial dependence on their spouses' earnings by the point of eligibility determination. The second "fairness problem involves the spouse who, though legally married at benefit time, has for years had no relationship with the covered worker. A brief marriage, long since ended practically but never legally, provides a basis for ultimate benefit entitlement.⁽⁹⁹⁾

Both cases involve a flow of benefits to individuals who have merely "won the marriage lottery." The losers include many others with far stronger "equitable" claims who have lived with covered workers over long periods, perhaps raised their children, and been financially dependent their covered earnings. To have entitlement hinge completely on marital status seems at times grossly unfair. The contrast is particularly poignant when a legal spouse (with a poor equitable claim) and another with much stronger emotional and financial ties to the worker, who cannot, however, meet the marital status test, claim against the same worker's account. The next two issues derive ultimately from this sense of unfairness.

B. Should Others – De Facto Spouses and Former Legal Spouses – Receive Spouse Benefits?

The basic marital test for spouse benefits looks to state law. The appropriate state is the covered worker's domicile at the time of the spouse's application or, for survivor benefits, at the time of the worker's death. A person qualifies if the courts of that state would find either:

(a) that the two were "validly married," or

(b) that the person claiming spouse benefits, though not legally married for some purposes, would "under the laws applied by such courts in determining the devolution of intestate personal property, have the same status with respect to the taking of such property as a [spouse of the worker].⁽¹⁰⁰⁾

This test permits the happenstance of local marriage and divorce law to control benefit entitlement under a national program. two people who have lived as husband and wife for thirty or forty years, under circumstances that would cause them to be "validly married" in some states or the equivalent of spouses for intestate succession purposes in others, can lose out on Social Security spouse benefits if they lived in the wrong place.⁽¹⁰¹⁾ Similarly, whether a first marriage, despite an attempted divorce and remarriage, leaves an eligible legal spouse depends on the state's divorce doctrines, including its recognition of foreign divorces.⁽¹⁰²⁾

Resolving these state law questions is often not easy. They intertwine with influential, if not dispositive, evidentiary presumptions (the presumptive validity of a second marriage, for instance) and procedural doctrines (such as estoppel) of dubious relevance in the Social Security context. Finally, sorting out a deceased wage earner's personal history can be difficult.⁽¹⁰³⁾

While preserving the basic test (with its attendant administrative difficulties), Congress has extended spouse benefits to two groups who do not meet it. In 1960, it established a purely "federal" marital status test that provides benefits to persons who have married in good faith, unaware of impediments to their marriages' validity.⁽¹⁰⁴⁾ This special provision, however, as enacted contained two provisos that the basic state law test did not include. First, such a "spouse" must be living with the wage earner at the time of death or application.⁽¹⁰⁵⁾ The second condition, since removed, was that there be no other spouse qualifying for benefits under the basic state law test.⁽¹⁰⁶⁾ (Since 1990 amendment, which removed this condition, a competing state law spouse and a spouse qualifying under the “deemed valid” marriage test can both receive benefits.  In cases where there are both types of spouse claimants, the state law spouse receives benefits outside the family maximum so that those benefits have no effect on the benefits received by the “deemed” spouse and any others whose benefits are subject to the maximum.)

In 1965, Congress extended benefits to a second group of non-spouses – namely, "former spouses" legally married to workers for a substantial period (initially twenty years, now ten by virtue of a 1977 amendment) but later divorced.⁽¹⁰⁷⁾ Providing benefits to former spouses, who no longer have either the requisite marital status or actual financial dependency at benefit time, represents an implicit but marked shift in justification for spouse benefits. Participation in a marriage, presumably as mother/homemaker, during a period when the covered worker built up a wage record necessarily replaces "presumed" need arising from the interruption of the working spouse's earnings. This change, more than any other, raises issues of how to deal with multiple spouses.

C. Should More Than One Spouse Receive Benefits on a Single Worker's Account?

The 1939 amendments limited the total of dependent benefits, including those to a spouse, that might be paid on the account of any contributing worker.⁽¹⁰⁸⁾ Further, the reference to state law for determining marital status assured that in most states there could be but one eligible spouse. Before Congress repealed it in 1957, the "living with" requirement combined with the "legal spouse" requirement virtually assured a limit of one spouse per worker. Today, the basic spouse eligibility test still limits most workers to one beneficiary spouse. In addition, the "family maximum" limits the total of monthly benefits paid on one worker's account at any one time. Depending on the worker's benefit level, the present "family maximum" is between 1.5 and 1.88 times the worker's basic benefit.⁽¹⁰⁹⁾ If the total amounts otherwise payable exceed the maximum, all benefits – with the exception of the worker's own retired-worker or disability benefits – are reduced proportionately to achieve that figure.⁽¹¹⁰⁾

The 1965 addition of divorced wives carried with it a complete break with these constraints. It introduced a significant possibility of at least two spouses qualifying on a single worker's account and removed the family maximum as a limitation. Benefits paid to a divorced spouse are not subject to the worker's family maximum and do not count against the maximum when it is applied to other benefits paid on the worker's account.⁽¹¹¹⁾ In 1977, Congress reduced the duration-of-marriage test for divorced wives to ten years, substantially increasing the extent to which the divorced-wife can undermine the family maximum. Theoretically, one worker may now leave a fair number of eligible divorced wives whose total claims, combined with those of diverse children and a current wife, far exceed the maximum. Since 1990 it has also been possible to throw in a wife or widow under an undissolved “deemed valid” marriage without impacting benefits subject to the maximum.

Because the present provisions still limit most wage earners to one eligible spouse, except where this is an eligible divorced wife, entitlement must sometimes be determined in an adversarial setting with competing widows (or occasionally wives) pitted against one another.⁽¹¹²⁾ The initial task in such cases is to determine the legal spouse under the appropriate law.⁽¹¹³⁾ In a rare case, state law may grant someone not a "legal wife" intestate succession rights, causing her to be a wife for Social Security purposes.⁽¹¹⁴⁾ This occasionally permits two "widows" to qualify for benefits on one wage account. as noted above, their benefits, along with those of any other eligible survivors, namely children, are subject to the worker's "family maximum."

Most states do not recognize a second marriage as valid when either of the parties has a prior marriage not legally dissolved.⁽¹¹⁵⁾ As a consequence, widow benefits go to a first wife who lived with the worker for a short period, rather than to a subsequent "wife" who spent many more years with him including those immediately preceding his death.⁽¹¹⁶⁾ The "good faith" spouse provision added to the Act in 1960, as modified in 1990, changes that result where the second “marriage” was entered in good faith and the potential beneficiary is living with the worker. ⁽¹¹⁷⁾

How these principles fit together and their ultimate arbitrariness are illustrated by the facts of a recent case.⁽¹¹⁸⁾ It concerned a wage earner, Samuel Hunter, and his four "wives." The district court summarized the relevant facts:

1) The deceased, Samuel Hunter, married Corinne prior to 1920. He left Corinne but was never divorced from her. Corinne died in 1950.

2) Samuel married Julia . . .on December 21, 1940. They lived together off and on until 1948 at which time [she] left Samuel after he threatened her and she never saw him or was in contact with him again. From 1940 to 1948 Julia had two or three children, one of whom was not Samuel's. There were never any divorce proceedings with respect to this marriage.

3) In 1944, Samuel married Rebecca but they separated at least eight or nine years before his death. There was no divorce.

4) In 1958, Samuel married again this time to Clara. He and Clara had had a child in 1953. They were living together at the time of Samuel's death.

5) Samuel died in 1964. All the events occurred in Louisiana.⁽¹¹⁹⁾

Although not the typical wage earner's personal history, it is simpler than some for purposes of calculating spouse benefits. Samuel at least had the decency to do all of his marrying, divorcing, and dying in one state.⁽¹²⁰⁾ Furthermore, he apparently "married" each of the four women ceremonially.⁽¹²¹⁾

Which of Samuel's four "wives" are eligible for widow benefits, assuming sufficient age (or age and disability) and low or nonexistent retirement benefits on their own accounts?

Corinne? Corinne did not survive Samuel. Had she survived him, assuming her marriage was valid under Louisiana law, she would qualify for widow benefits. The duration of her marriage (being more than nine months before Samuel's death) is irrelevant, as is the fact that other wives stayed with Samuel longer or bore him children. The long separation, the subsequent "marriages," that she neither lived with Samuel nor was supported by him at the time of his death – none of these facts would have frustrated Corinne's claim had she survived Samuel.

Julia? Samuel was married to Corinne at the time he "wed" Julia. Under Louisiana law, therefore, Julia and Samuel were not validly married, and Julia is not Samuel's widow.

The alternate state law test is a possibility because Louisiana allows "putative spouses" intestate succession rights equivalent to those of the legal spouse. In this case, however, the Social Security Administration determined, on the basis of contradictory testimony, that Julia did not meet the "good faith" requirement of the Louisiana putative spouse doctrine.⁽¹²²⁾

The "federal" marital status test also requires "good faith." Although conceivable more lenient than Louisiana's stand, the federal test cannot help Julia, for she was not living with Samuel at the time of his death.

Rebecca? Rebecca is in the same situation as Julia. If she meets the "good faith" test of the Louisiana putative spouse doctrine she can qualify for widow benefits; it is her only chance.⁽¹²³⁾

Clara? By the time Samuel married Clara, Corinne had died. Consequently, Samuel and Clara were validly married under Louisiana law. Even if Corinne were alive when Samuel married Clara, Clara might still have qualified as a putative spouse. Clara might also have met the "federal" good faith spouse test, for she alone lived with Samuel at the time of his death. But this test would require her to prove good faith.⁽¹²⁴⁾

To conclude, as the system now stands, a woman who lived for years as the "wife" of a covered worker and depended upon his earnings as her only significant source of income, may not qualify for spouse benefits. Her eligibility rests on such accidents as where the two have lived and whether there is another woman, perhaps unknown to her, still legally married to the worker. If such a woman exists, further accidents, including that woman's age and Social Security earnings history, govern the later "wife's" entitlement.

One can imagine improvements to the system that would make spouse-benefit entitlement less accidental, but how much less? The worker's insured status and basic benefit award usually rest on a lengthy period of work. During that time there may have been several spouses, legal and de facto. The inescapable dilemma is how to allocate spouse benefits among them. The original eligibility test expressed a clear choice. In specifying "legal spouse, living with or supported by the worker," it directed benefits to an individual likely to suffer income loss upon the worker's death, retirement, or disability. The abandonment of that principle leaves no compelling basis for choosing who among several "spouses" will receive benefits – assuming Congress retains an essentially all-or-nothing system. Nor does any administratively attractive way exist to divide spouse benefits among them. Finally, giving more than one of them full spouse benefits, currently allowed in the case of divorced ten-year wives, raises a serious question about the system's fairness to contributors with only one eligible spouse or none.

D. The Remarriage Dilemma

The remarriage dilemma poses a final sample of the inherent arbitrariness of spouse benefits. Initially, benefits for surviving spouses, resting on the interruption of earnings on which the system presumed the individual to have been dependent, terminated upon remarriage.⁽¹²⁵⁾ Remarriage, so the logic went, supplanted the need for benefits with a new dependent relationship. Consistent with that rationale, the 1958 Social Security Act Amendments created a complex set of exceptions to avoid terminating dependent or survivor benefits when two such secondary beneficiaries married each other – the marriage of a person receiving widow benefits to a person receiving widower benefits, for example.⁽¹²⁶⁾

That exception, however, did not significantly ameliorate the most troublesome feature of the remarriage provision – the financial loss it attached to the legal marriage of a retired male worker and a widow collecting benefits on the account of a deceased husband. This potential loss reportedly caused many elderly couples to "live in sin."⁽¹²⁷⁾ Concern over the perceived inequity of treating married elderly couples less favorably than those cohabiting without marriage and the resulting discouragement of legal marriage created strong pressure for the removal or reduction of the remarriage penalty.

A major problem of equity, however, lay on the other side. Without loss or adjustment of widow benefits upon remarriage, two couples, otherwise identical, could receive substantially different benefits. The first, a retired-worker husband and his dependent wife, would receive one and one-half times his primary insurance amount. The second, a retired-worker husband and his wife, the widow of another (identical) worker, would receive more (two full primary insurance amounts under the current widow-benefit percentage).

Continued attention to the "living in sin" problem led Congress, in 1965, to permit remarriage from sixty on for widows (sixty-two for widowers) with only a fifty percent reduction in benefits (based on the original spouse's primary insurance amount) rather than a complete loss.⁽¹²⁸⁾ Then in 1977, Congress allowed remarriage for this age group without any effect on benefits.⁽¹²⁹⁾ This solution inescapably resulted in a situation in which widow/retired-worker and widower/retired-worker married couples receive "more than other couples would get where the husbands [or wives] had an identical record of covered earnings."⁽¹³⁰⁾

Inequity on one side or another is simply unavoidable; it inheres in a system of benefits tied neither to actual need nor to direct or derivative contribution.

VI. Improving Spouse Benefits as a Technique For Improving Benefits for Women

When benefit improvement did not carry politically troubling tax consequences, Congress quite naturally responded to charges of "arbitrary" benefit denial by expanding the eligible class.⁽¹³¹⁾ Although those times are past, Congress continued to view the spouse-benefit structure as a device for improving the relative benefit status of women. Examples include the 1977 amendments which remove the remarriage penalty on spouse benefits after age 60 and reduced the divorced spouse durational test to 10 years.

Without question, such provisions can improve the total payout to women. Many more women than men receive spouse benefits. Accordingly, manipulation of the spouse-benefit system can easily achieve greater equality in gross – more nearly equal average payments for women. Moving from averages to individuals, however, reveals the deficiency of such measures. They do no more than pump additional revenue through a fundamentally inequitable system.

The preceding section explored some of the inequities on the benefit side. More troublesome still, in light of ever-rising Social Security taxes, are the inequities imposed on those contributors who reap no advantage whatsoever – either directly or through a spouse – from spouse benefits. Reforms that improve the value or availability of spouse benefits either add to those inequities or, at a minimum, shift their impact.

The Social Security tax is no longer a trifle. Discrepancies between contribution and return, disregarded at much lower tax rates, have become economically and politically important. Under these circumstances, a discrepancy biased against a growing majority of the contributing workforce is particularly troubling. Such is the case with spouse benefits. social security taxes workers, without regard to marital status, at a very high rate. The system pays handsome bonuses on behalf of some contributors legally married at the time of death, disability, or retirement. The terms of those extra benefits give preferential treatment to marriages that endure (including 10-year former marriages) and to marriages in which one spouse has a Social Security earnings record while the other has none or one much smaller.

Once the national norm, that preferred group grows less typical each year. First, over much of their course, most marriages today are two-worker not one-worker marriages. Over half the married couples below normal retirement age now report earnings from both spouses.⁽¹³²⁾ Second, many marriages no longer endure from the parties' entry into the labor force until retirement. In 1940, when wife and widow benefits began, the ratio of divorces to marriages was about 1 to 6.⁽¹⁵⁰⁾ Currently, it is more nearly 1 to 2.⁽¹³³⁾ Of the generation currently receiving Social Security, only fourteen percent of women ever married saw their first marriage end in divorce.⁽¹³⁴⁾ For the present generation of young women, evidence indicates that as many as one-third "may eventually end their first marriage in divorce."⁽¹³⁵⁾ And the heaviest incidence of divorce precedes the 10-year threshold set by the 1977 amendments.⁽¹³⁶⁾ Many of those approaching retirement age are currently divorced or separated from their most recent spouse.⁽¹³⁷⁾ Third, increasing numbers of workers are single, the result of the higher divorce rate and a greater initial reluctance to marry.⁽¹³⁸⁾

Despite the inequity and administrative difficulty associated with spouse benefits, abolition standing alone is not a feasible or desirable reform. Notwithstanding the increasing number of women eligible for retired-worker benefits and the added support of spouse benefits, women far less well under the present system than do men. they outnumber men among all Social Security beneficiaries (as they do among the elderly population), but they receive dramatically lower average benefits (see Table 1).⁽¹³⁹⁾ Nearly half of these benefits are secondary (see Figure 1). Elimination of wife and widow benefits would exacerbate the present sex disparity.

VII. An Alternative: Recognition of Marriage on the Contribution Side

A. Retired-Worker Benefits and Women

An acceptable phase-out of spouse benefits requires simultaneous alteration of the method of calculating retired-worker benefits. Retired-worker benefits depend on average earnings during periods of covered employment. Although wage parity between men and women may lie ahead, in the near future women as a group will continue to receive lower Social Security retired-worker benefits because of lower wages.⁽¹⁴⁰⁾ There is partial mitigation in the system's "tilt"⁽¹⁴¹⁾ – lower wage employees receive a better return in Social Security benefits than higher wage employees – but not enough to offset the wage disparity. Since more men than women earn above the taxable wage base for Social Security, raising the wage limit, as Congress did in 1977, increases the sex differential.⁽¹⁴²⁾

The treatment of years spent out of covered employment also affects the benefits retired women receive from Social Security. In the past, the work lives of women – in contrast to those of men – have incorporated significant periods out of the workforce, typically spent in the home caring for children.⁽¹⁴³⁾ The retired-worker benefit formula ignores up to five post-1950 years in calculating average covered wages.⁽¹⁴⁴⁾ Increasing that number substantially or ignoring all years of work at home would yield higher retired-worker benefits for a great many women.⁽¹⁴⁵⁾ A step beyond ignoring years spent at home would be to treat them as producing credits toward eventual benefits-in old age or disability. A few have proposed, in this vein, that a constructive wage for Social Security purposes should attach to work in the home,⁽¹⁴⁶⁾ but serious practical problems and theoretical objections surround such an approach.

B. The Community-Property Model

A far more promising reform, which would both yield credits for years in the home and tend to offset the male-female wage differential, derives from an existing legal model-community property. Applying the community-property concept to Social Security would require splitting earnings credits between spouses as they are created. Married men and women would accumulate equal wage records regardless of their respective wage rates or whether one devoted substantial or full time to nonmarket, "community" affairs: cleaning, meal preparation, or child rearing. Marital status would be taken into account from year to year on the contribution side, rather than only a certain critical points or periods s it is under the current spouse-benefit system. Credits received as spouse and as worker would cumulate, not offset one another. Credits would also remain secure through divorce and entry into or exit from the labor force. A variant on this approach would split records over the period of a marriage, retrospectively, upon divorce.

The community-property states⁽¹⁴⁷⁾ have long treated married couples, legally, as economic partnerships, presuming that income earned by either spouse is the product of a joint effort. As a result, earnings belong to both spouses in equal shares.⁽¹⁴⁸⁾ Since 1948, federal income tax law has contained a comparable principle within the joint return, which permits income splitting between husband and wife.⁽¹⁴⁹⁾ The marital deduction in the federal estate and gift tax reflects the same pattern.⁽¹⁵⁰⁾ Similarly, divorce law reform in many noncommunity-property states has brought the notion of equal shares or at least "equitable division" to the dissolution of marriages.⁽¹⁵¹⁾

Although applying community-property doctrine to retirement plans has proven difficult,⁽¹⁵²⁾ logic and practical considerations have compelled the courts in community-property states to take that step.

Over the past decades, pension benefits have become an increasingly significant part of the consideration earned by the employee for his services. As the date of vesting and retirement approaches, the value of the pension right grows until it often represents the most important asset of the martial community. . . . A division of community property which awards one spouse the entire value of this asset, without any offsetting award to the other spouse, does not represent that equal division of community property contemplated by [the statute].⁽¹⁵³⁾

All community-property states hold that pension rights resting in part on employment by one of the spouses during marriage are community property.⁽¹⁵⁴⁾ The issue most often arises during division of community property upon divorce. It is likely also to arise in cases of disagreement over the exercise of pension options in states that provide for joint management of community assets.⁽¹⁵⁵⁾

C. Earnings-Splitting Applied to Social Security

Ample reasons exist for Congress to retain full control over the allocation of Social Security benefits. The suggestion advanced is not that Social Security be made subject to state community-property or divorce law, but rather that Congress should itself rework the system to internalize the community-property or marital-property concept. This would align the program with current legal and social trends,⁽¹⁶³⁾ while avoiding most of the practical problems that have accompanied state court efforts to divide pensions between spouses. Those difficulties derive in no small part from the disparity between the judicially applied "community" concept and the terms of most pension plans, which treat benefits and the exercise of benefit options as belonging exclusively to the worker. Social Security, as an essentially universal national program, can feasibly avoid this tension by incorporating the community-property approach directly within its benefit structure. The change would, of course, pose difficulties of transition, as do nearly all significant Social Security benefit revisions. But the resulting scheme would be fairer and easier to administer than the present dual system of worker benefits and supplemental spouse benefits.

In the 95^th Congress, Representatives Donald Fraser and Martha Keys introduced a bill that would have split Social Security earnings records between spouses who elected to file a joint federal income tax return.⁽¹⁶⁴⁾ While the bill had gained fifty-eight sponsors and the support of the National Organization for Women, it did not receive serious attention from Congress.⁽¹⁶⁵⁾

The proposal envisioned the eventual elimination of spouse benefits, including those for widows and widowers. Instead, each spouse would have possessed an independent entitlement to old-age and disability benefits. Periods without paid employment that a spouse devoted to child rearing, other work in the home, or education would not have left blank years for Social Security if the parties chose to report split earnings. Discrepancies between sexes in work patterns and wage rates would, upon spousal election, have balanced out. Divorce or separation would have not threatened benefits for they would not any longer have depended on marital status at benefit time. A divorced or separated spouse would have taken from the marriage a wage record reflecting his or her share of its earnings.

Under the Fraser-Keys bill housewives (or househusbands) becoming totally disabled would, in most cases, have had sufficient recent wage entries to qualify for disability benefits. Under the present system, a brief period out of covered work, including a period of work in the home, causes loss of the insured status necessary for disability benefits.⁽¹⁶⁶⁾

Aside from the complexities of transition,⁽¹⁶⁷⁾ such a proposal to split spousal earnings, whether year-by-year or only upon divorce, poses several difficult issues. One spouse, for example, may earn well above the contribution base, the other far below it. Should the split bring more of the high income spouse's earnings under the tax? The Fraser-Keys bill provided only for splitting the amount that, on an individual basis, fell under the taxable wage ceiling.⁽¹⁶⁸⁾

The system's present benefit bias toward enduring one-earner marriages, vis-à-vis single persons and couples with comparable earnings histories, poses a potential political problem of larger dimensions. Straight income splitting would likely mean benefit reduction for that currently preferred group. The Fraser-Keys bill would have retained the present preference by allowing couples who filed joint returns to receive individual wage credits equal to fifty percent of their join wages or seventh-five percent of the wages of the higher paid spouse, if that was more.⁽¹⁶⁹⁾ The latter option assured the enduring one-earner marriage of retired-worker benefits based on 150% of the average wage of the high income spouse, just as the present system yields spouse benefits of fifty percent in addition to the worker's retirement benefit.

The final issue raised by a year-by-year earnings-splitting approach involves marriages in which one spouse, typically the spouse with higher earnings (the husband), is several years older than the other. Alicia Munnell puts the problem quite concisely:

Consider the case of a male worker ages sixty-five who wishes to retire but whose nonworking wife is too young to collect her benefits. With the mandatory division [of credits], the husband can collect only half the benefits the couple is entitled to and thus may not be able to afford retirement until [his wife is older].⁽¹⁷⁰⁾

This situation poses difficulty for earnings-splitting schemes, but the problem is not, as a few studies would imply, insurmountable.⁽¹⁷¹⁾ Congress could lower the retirement age for the younger low-earnings spouse or offer an optional allocation of more than half the enduring couple's combined credits to the older spouse. Workable devices for alleviating the hardship are far from unimaginable.⁽¹⁷²⁾ Conversion to earnings-splitting would no doubt require coordinated changes in other features of the Social Security benefits system; it would, in fact, re-form that system. The need for such ancillary changes is not a legitimate ground for rejecting the approach if the proposed system promises fairer recognition of both earnings add marriage and if the transition from the present worker/spouse dual-benefit approach is administratively and politically feasible.

Earnings-splitting does not necessarily remove from Social Security difficult questions of fact and law bearing on whether two persons are "validly married." If constructed carefully, however, conversion would drastically reduce the importance of such questions. The Internal Revenue Service, for example, primarily relies on the parties' own declaration of status in determining eligibility to file a joint return.⁽¹⁷³⁾ The current Social Security system stresses such questions because so much – eligibility for spouses benefits – rides on the answer.

A system constructed without the current heavy bias in favor of couples with widely disparate incomes might allow splitting on an optional basis, without regard to marital or family connection. Any two persons might split their wage credits. Tying the election to the federal income tax return, however, as the Fraser-Keys bill did, is not a bad solution. It is particularly attractive if the marital status on which any year's split record rests will, after a short period, enjoy a strong statutory presumption of validity.⁽¹⁷⁴⁾

Conclusion

Spouse benefits responded quite reasonably to the mores as well as the practicalities of another era. They succeeded in fulfilling the purpose that led Congress to append them to the system in 1939, without serious inequity to contributors and recipients or waste to the system. In recent decades, however, dramatic social, economic and legal changes have altered the profile of the American family, at home and at work. Since 1939, the Social Security system itself has grown and changed enormously. In today's environment, the current system – primary benefits based on the worker's prior employment set off against secondary benefits based on marriage to a covered worker – functions awkwardly. Its allocation of non-contributory spouse benefits seems quite arbitrary, and the tax burden associated with those benefits weighs heavily on the large and growing portion of the work force that does not receive them. Moreover, the system's failure to produce benefits for women comparable to those it yields for men can no longer be lightly dismissed. To respond adequately to these developments, Congress would have to revise the Social Security benefit scheme substantially. Among available reforms, the incorporation of the community-property concept in the form of an equal division of earnings between spouses, year by year as they are obtained, seems the most attractive.

---

Endnotes

* A much earlier version of this essay appeared in 63 Cornell Law Review 789 (1978). Copyright © 1978, 2005 Peter W. Martin.

1. United States Justice Department Task Force on Sex Discrimination, Staff memorandum 10 (Mar. 23, 1977). This memorandum figured prominently in the statement of Rep. Donald M. Fraser against the administration's 1977 dependent-benefit proposals. See President Carter's Social Security Proposals: Hearings Before the Subcomm. on Social Security of the House Comm. on Ways and Means, 95^th Cong., 1^st Sess. 572 (1977) [hereinafter cited as 1977 Hearings].

2. 123 Cong. Rec. S19,505 (daily ed. Dec. 15, 1977).

3. See HEW, Report of the Task Force on the Treatment of Women Under Social Security 11-12 (1978) [hereinafter cited as HEW Task Force Report]:

In 1975, 93 percent of elderly men and 51 percent of elderly women received social security beneits as retired workers. . . . The women include 11 percent who received higher dependents' benefits in addition to their retired-worker benefits and 40 percent who received only retired-worker benefits. Most of the remaining 49 percent of elderly women were entitled to benefits as dependent wives or widows.

4. Social Security Administration, Annual Statistical Supplement, 2003, Table 5.G2.

5. See < http://www.ssa.gov/pressoffice/factsheets/women.htm >.

6. National Economic Council Interagency Working Group on Social Security, Women and Retirement Security 11 (1998).

7. The amount paid to the worker's estate, like the monthly retired-worker benefit, was based upon the total of covered wages on which contribution (tax) had been paid. The Act directed payment of a full 3.5% of such total covered wages to the estate of a worker who died before collecting old age benefits; 3.5% minus the total of monthly payments received before death, to the estate of a deceased retired worker. Social Security Act of 1935, ch. 531, § 203, 49 Stat. 620. Compared to the present law, the simplicity of the 1935 legislation is overwhelming: the Old Age Insurance provisions took up less than four pages. For a good summary, see Social Security Board, Social Security in America 222-26 (1937).

8. The 1939 amendments, with few exceptions, embodied the recommendations contained in a report of the Social Security Board. See Social Security Board, Proposed Changes in the Social Security Act (1939) [hereinafter cited as Soc. Sec. Bd. Proposed Changes], reprinted in Social Security: Hearings Relative to the Social Security Act Amendments of 1939 Before the House Comm. on Ways and Means, 76^th Cong., 1^st Sess. 3 (1939) [hereinafter cited as 1939 Hearings}; Social Security Act Amendments of 1939, ch. 666, 53 Stat. 1360.

The Board's recommendations, in turn, closely tracked those of a 1938 Advisory Council on Social Security. See Advisory Council on Social Security, Final Report, S. Doc. No. 4, 76^th Cong., 1^st Sess. (1938), reprinted in 1939 Hearings, supra at 18.

9. The shift in the retired-worker benefit formula – from a total contribution basis to an average covered-wages basis – had the effect of improving immediate payout by relaxing the benefit-contribution relationship. See Soc. Sec. Bd. Proposed Changes, supra note 8, at 5-6. See generally A. Altmeyer, the Formative Years of Social Security 96-117 (1966).

10. Soc. Sec. Bd. Proposed Changes, supra note 8, at 4. Although eligible retired workers could be of either sex, these spouse benefits were exclusively for "aged wives."

11. The Social Security Board's report argued for the substitution in these terms:

Under a social insurance system the primary purpose should be to pay benefits in accordance with the presumptive needs of the beneficiaries, rather than to make payments to the estate of a deceased employee regardless of whether or not he leaves dependents. The payment of monthly benefits to widows and orphans, who are the two chief classes of dependent survivors, would furnish much more significant protection than does the payment of lump-sum benefits.

Id. at 6.

In addition to providing for wives and widows, the amendments created dependent and survivor benefits for children of a contributor and survivor benefits for financially dependent elderly parents (the latter only if no wife or eligible children survived). Social Security Act Amendments of 1939, ch. 666, § 201, 53 Stat. 1360 (adding § 202(c), (f)).

12. Soc. Sec. Bd. Proposed Changes, supra note 8, at 4. The Senate Finance Committee expressed the same view. See S. Rep. No. 734, 76^th Cong., 1^st Sess. 11 (1939) (accompanying H.R. 6635), See also 1939 Hearings, supra note 8, at 1014, 1218.

13. The 1939 amendments halted a scheduled increase to 1.5%. See Social Security Act of 1935, ch. 531, § 801, 49, Stat. 620; Social Security Act Amendments of 1939, ch. 666, § 601, 53 Stat. 1360. Congress similarly deferred subsequent scheduled tax increases until 1950, when the tax rate rose to 1.5%. See Social Security Act Amendments of 1950, ch. 809, § 201, 64 Stat. 477 (now I.R.C. § 3101).

14. The original legislation provided greater payments – in relation to contribution – for beneficiaries with low wages and few years of covered employment before retirement. See Social Security Act of 1935, ch. 531, § 202(a), 49 Stat. 620.

15. Soc. Sec. Bd. Proposed Changes, supra note 8, at 2.

16. The Board put it thus: "In order that greater social adequacy may not be achieved at the expense of individual equity, the Board recommendations that the benefits payable to unmarried persons continue to be at least as much as they could purchase from a commercial insurance company with their own contributions." Id. at 4-5.

17. That maximum was (a) twice the wage earner's primary insurance amount, (b) 80% of his average monthly wage, or (c) $85, whichever was least. Social Security Act Amendments of 1939, ch. 666, § 201, 53 Stat. 1360 (adding § 203(a)).

18. Soc. Sec. Bd. Proposed Changes, supra note 8, at 12.

19. Of course, for the first wave of retirees, marriage during the years immediately prior to retirement did correspond to the period of covered employment. To qualify as a wife at 65, a woman had to be married to the worker before he turned 60. Shorter marriage