The Never-Married in Old Age: Projections and Concerns for the Near Future

by
Social Security Bulletin, Vol. 67, No. 2, 2007

Related Content

Population Fact Sheet: Marital Status & Poverty

This article focuses on a growing yet understudied subgroup of the elderly in the United States-the never-married. The first section, based on data from the Current Population Survey and a review of the academic literature, examines the current circumstances of never-married retirees, particularly their economic and health well-being. The succeeding section uses the Modeling Income in the Near Term (MINT) model to assess the projected (1) changes in the marital status composition of the future retirement-age population; (2) demographics of future never-married retirees, and (3) economic well-being of never-married retirees. The results highlight important links between marital trends, Social Security, and retirement outcomes and offer insight into some of the characteristics of current and future never-married retirees.


The author is with the Office of Retirement Policy, Office of Policy, Social Security Administration.

Acknowledgments: The author thanks David Weaver for suggestions on earlier versions of this paper. The author also thanks reviewers and the editor for helpful comments. Mark Sarney and other MINT data users provided exceptional research support.

The findings and conclusions presented in the Bulletin are those of the authors and do not necessarily represent the views of the Social Security Administration.

Summary and Introduction

Understanding how marital patterns affect the Social Security program and its beneficiaries has become an important policy and academic focus. Over the past several decades, a convergence of economic, demographic, and social changes has given rise to dramatic changes in marital trends in the United States. Divorce rates surged in the 1960s and 1970s, the age of first marriage has risen, and the number of persons never marrying has increased (Ruggles 1997; Goldstein 1999; Goldstein and Kenney 2001; Kreider 2005; Harrington Meyer, Wolf, and Himes 2006). Evidence also indicates that the remarriage rate has decreased, and dissolution of second marriages has risen (Cherlin 1992; Norton and Miller 1992). Put together, these trends suggest that a rising share of unmarried people will be entering retirement age in the near future.1

A growing body of economic, sociological, and demographic research has highlighted an association between marital status and adult well-being. A moderate-to-strong relationship has been found between marital status and an individual's economic resources (Waite and Gallagher 2000; Wilmonth and Koso 2002) as well as health profile (Schoenborn 2004). However, a comparatively small amount of the literature has focused specifically on the elderly population and differences among the unmarried elderly—individuals who have never married or are divorced or widowed—are even less examined. Among the unmarried, women who are widowed (Morgan 1992; Weaver 2002) or divorced (Weaver 1997; Butrica and Iams 2000) have received the majority of attention, while the never-married are often overlooked.

This article focuses on a growing yet understudied subgroup of the elderly in the United States: the never-married, meaning persons who have never been legally married or whose marriages ended in annulment. Its purpose is to assess how never-married persons fare during retirement—at present and as the large baby-boom generation retires.

Although never-married retirees are not typical Social Security beneficiaries, they are by no means an insignificant population. In 2003, about 4 percent of Americans aged 65 or older, or 1.4 million individuals, had never married (He and others 2005, Table 6.1). Moreover, the share of retirement-age persons who have never married is projected to increase as the baby-boom cohort reaches retirement age (Easterlin, Schaeffer, and Macunovich 1993, 508–509; Butrica and Iams 2000, Table 1; Harrington Meyer, Wolf, and Himes 2004). The Urban Institute's DYNASIM3 model, for example, predicts that never-married persons will increase to around 6 percent of the retirement-age population by 2040 (Favreault and Smith 2004).

The projected growth of never-married retirees raises a number of important issues for retirement policy. Because Social Security spousal and survivor benefits are determined by marital history, changes in marital trends can have important implications for Social Security program costs and distributional outcomes among its beneficiaries.2 A rise in the share of persons entering retirement as never-married would, for example, contribute to a decline in individuals eligible to receive auxiliary benefits and, correspondingly, a rise in beneficiaries receiving only retired-worker benefits (see, for example, Harrington Meyer, Wolf, and Himes 2006). Another issue relates to the economic well-being of retirees. Although Social Security reform plans have given great attention to widows because of their greater likelihood of economic insecurity in old age (Weaver 2002), the never-married may also tend to experience a heightened risk of economic hardship in retirement.

The first section of the article, based on data from the Current Population Survey and a review of the academic literature, examines the current circumstances of never-married retirees, particularly characteristics of their economic and health well-being. The next section shifts focus to the near future. Using the Social Security Administration's (SSA's) Modeling Income in the Near Term (MINT) model, demographic and economic projections of the population aged 62 or older are assessed for the years 2020, 2030, 2040, 2050, and 2060. These data are exceptionally useful for analyzing and projecting changes in the marital status composition of the population at retirement age, the demographics of future never-married retirees, and economic well-being (poverty rate, income distribution, and welfare ratio) of never-married retirees.

The results highlight important links between marital trends, Social Security, and retirement outcomes. Although the never-married represent an economically diverse group, poverty among the elderly who have never married is particularly high—more than twice the national average in 2004, four times higher than that of married persons, and greater than the poverty rates of the divorced and widowed. In addition, a review of existing studies suggests that never-married persons are more likely to have health risks during retirement that are greater than those of married persons and the national average. MINT projections indicate important changes in the marital composition of future retirees marked by a rising proportion of never-married persons entering retirement age. Future never-married retirees, according to MINT, are expected to have the highest elderly poverty rate among marital groups.

Finally, this study calls attention to heterogeneity among the never-married elderly. At least two very different population segments exist among the never-married: one with greater than average economic resources and educational attainment and another with little economic resources and educational attainment. Thus, it may be important to look at the never-married in greater detail.

Marital Status and Retirement Risks, with Emphasis on the Never-Married

Family structure shapes retirement experience. Research suggests that unmarried older adults are generally at a disadvantage compared with married persons, in terms of economic security but also in health areas (Lillard and Panis 1996; Wilmonth and Koso 2002; Keith 2003). In "Does Marriage Matter?" sociologist Linda Waite (1995) dubs this the marriage "advantage," which intends to underscore the multiple benefits of marriage for adult well-being. Along this line, existing evidence indicates that marital status can influence retirement timing (Morgan 1992; Gustman and Steinmeier 2000; Pienta and Hayward 2002), the economic resources available to older adults (Butrica and Iams 2000), as well as their health condition (Lillard and Panis 1996). Marital history also determines eligibility for Social Security benefits for spouses and survivors, which can represent an important source of retirement income, especially for widowed women.

With that said, our understanding of the links between marital status and retirement outcomes remains limited. Much of the research is based on samples of the working-age population. Studies that do focus on older adults tend to lump the unmarried (widowed, divorced, and never-married) into a single category.3 Evaluating the unmarried as a whole can obscure important differences between the widowed, divorced, and never-married. Important exceptions include work on widowed and divorced women (Morgan 1992; Butrica and Iams 2000; Weaver 2002).

Although frequently overlooked in policy and academic discussions, never-married retirees make up a noteworthy share of the U.S. retirement-age population (Table 1). According to U.S. census figures in 2000, around 4 percent of men and women aged 65 or older had never married. Meanwhile, the share of never-married men and women among the 45–64 age group was roughly double (8.1 percent and 6.8 percent, respectively). Although these figures are important, they do not tell us how never-married persons fare in retirement. To begin addressing this issue, empirical evidence about the never-married is assessed across two dimensions of well-being: economic and health. Doing so provides a more complete portrait of the never-married than is typically presented in the literature.

Table 1. Percentage distribution of adults in the United States aged 45 or older, 2000, by age, sex, and marital status (in percent unless otherwise noted)
Sex and age Total Now-
married a
Never-
married
Divorced Separated b Widowed
Number Percent
Total 45 or older 96,728,811 100.0 64.6 6.3 12.6 2.0 14.5
Aged 45–64
Men, 45–64 29,994,964 100.0 74.1 8.1 13.9 2.2 1.7
45–54 18,425,577 100.0 72.2 9.7 14.7 2.4 1.0
55–64 11,569,387 100.0 77.0 5.6 12.6 1.9 2.8
Women, 45–64 31,754,875 100.0 66.1 6.8 17.3 2.8 7.0
45–54 19,153,032 100.0 67.1 8.0 18.0 3.1 3.7
55–64 12,601,843 100.0 64.5 5.0 16.3 2.3 11.9
Aged 65 or older
Men, 65 or older 14,382,370 100.0 73.8 4.4 6.7 1.2 13.9
65–74 8,355,575 100.0 77.4 4.6 8.3 1.4 8.3
75–84 4,823,419 100.0 71.9 4.1 4.9 0.9 18.2
85 or older 1,203,376 100.0 56.3 4.3 3.3 0.8 35.3
Women, 65 or older 20,596,602 100.0 41.9 4.3 7.5 1.0 45.3
65–74 10,145,574 100.0 53.7 4.1 10.1 1.3 30.8
75–84 7,493,843 100.0 34.8 4.3 5.8 0.7 54.6
85 or older 2,957,185 100.0 19.4 5.2 3.3 0.5 71.6
SOURCE: Author's calculations using U.S. Census Bureau 2000 data, presented in Kreider and Simmons (2003), Table 1.
a. Includes spouses present and absent.
b. Includes people who were not living with their spouse because of marital discord.

The Economic Dimension

Poverty. Research suggests that unmarried persons are more likely to face prospects of lower income in retirement than married individuals. One telling measure of an association between marital status and economic outcomes in old age is the incidence of elderly poverty by marital group. Notably, the never-married have the largest share of persons aged 65 or older in poverty (21.9 percent) compared with 4.5 percent of married persons, 14.5 percent of widowed persons, 17.3 percent of divorced persons, and 9.8 percent overall (Chart 1). That is to say, the elderly poverty rate among the never-married is more than four times the married rate, more than double the national average, and greater than the rates of other unmarried groups. To assess statistical differences across marital categories, a test of differences based on the square root of the sum of the squares of the standard errors was calculated. For the poverty measure, results show a statistically significant difference (at the 0.05 level) between the poverty rate of the never-married aged 65 or older and that of all other marital groups.4

Elderly poverty rates differ not only across marital subgroups but also by sex (Chart 1).5 Particularly striking, the poverty rate for elderly never-married women was more than four times that of their elderly married counterparts as of 2004 (21.3 percent and 4.4 percent, respectively); and among elderly unmarried women, a larger share of never-married women experienced poverty (21.3 percent) compared with the shares of divorced (20.7 percent) and widowed (15.4 percent). Tests indicate statistically significant differences (at the 0.05 level) between never-married, widowed, and married women on the poverty measure but no significant difference between never-married and divorced women. For elderly men, the never-married had the highest prevalence of poverty, at 22.6 percent—far higher than that of their married (4.6 percent) and unmarried counterparts (divorced men, 12 percent, and widowed men, 10.9 percent). These differences are statistically significant at the 0.05 level. Thus, it is not just unmarried women who are at risk of old-age poverty but also never-married men.

Chart 1.
Percentage of persons aged 65 or older below the poverty line, by marital status and sex, 2004
By marital statusa
Bar chart with tabular version below.
By marital status and sexb
Bar chart with tabular version below.
Show as table
Table equivalent for Chart 1. Percentage of persons aged 65 or older below the poverty line, by marital status and sex, 2004
Status Percent by
marital status a
Percent by marital status and sex b
Men Women
All 9.80 7.00 12.00
Married 4.50 4.60 4.40
Never-married 21.90 22.60 21.30
Divorced 17.30 12.00 20.70
Widowed 14.50 10.90 15.40
 
SOURCE: Author's calculations using data from Social Security Administration (2006b, Table 8.1). Data are based on the U.S. Census Bureau's Current Population Survey, March Supplement.
NOTES: Poverty rates are based on total income of the family—sum of total money income of all persons related by blood, marriage, or adoption and residing together—compared with official poverty thresholds of elderly families in 2004. Total money income includes all income regularly received by the family before any deductions, including wages, Social Security, Supplemental Security Income, public assistance, interest dividends, rent, royalties, and veterans' payments. Calculations do not include nonmoney income transfers, such as food stamps, health benefits, or subsidized housing. Income refers to receipts for calendar year 2004, whereas marital status refers to the date of the survey.
Persons who are separated or married but living apart are included in "All" but are not shown separately.
a. Population totals (in thousands) are 35,213 (All), 19,278 (Married), 1,460 (Never-married), 2,777 (Divorced), and 10,682 (Widowed).
b. Population totals (in thousands) for men are 15,151 (All), 10,858 (Married), 670 (Never-married), 1,070 (Divorced), and 2,069 (Widowed). For women, the totals are 20,063 (All), 8,420 (Married), 790 (Never-married), 1,707 (Divorced), and 8,613 (Widowed).

To evaluate the historical evolution of elderly poverty across marital groups, tabulations were compiled using various years of the U.S. Census Bureau's Current Population Survey, March Supplements. The results show that the overall poverty rate of the elderly fell quite dramatically from 1970 to 2004 (Chart 2). Part of this decline can be attributed to general increases in Social Security benefits during the 1970s, along with other changes in the program. Underlying the dramatic drop in poverty among widow(er)s in the 1970s, for example, was legislation that augmented the survivor benefit rate from 82.5 percent of a deceased spouse's primary insurance amount to 100 percent (Martin and Weaver 2005, 8). Another example is the rule concerning the length of marriage, which in 1977 reduced the number of years required for receipt of benefits for divorced spouses and divorced survivors from 20 years to 10 years.

Chart 2.
Poverty rates of persons aged 65 or older, by marital status, 1970–2004
Line chart with tabular version below.
Show as table
Table equivalent for Chart 2. Poverty rates of persons aged 65 or older, by marital status, 1970–2004
Year All Never-married Divorced Widowed Married a
1970 24.43 29.23 32.53 34.72 15.50
1975 15.47 22.83 23.95 22.80 8.50
1980 15.76 21.40 27.20 24.13 8.30
1985 12.74 20.50 20.60 20.30 5.80
1990 12.15 18.10 21.20 20.00 5.10
1995 10.48 25.90 15.40 17.40 3.80
2000 10.19 25.60 16.70 15.90 4.50
2004 9.81 21.90 17.30 14.50 4.50
 
SOURCE: U.S. Census Bureau, Current Population Survey, March Supplement (1971, 1976, 1981, 1986, 1991, 1996, 2001, 2005).
a. Does not include persons who are married but living apart from their spouse.

At the same time, the data show that the degree of decline of elderly poverty between 1970 and 2004 differs by marital group, with the never-married rate decreasing less dramatically than those of other marital groups. For married individuals aged 65 or older, poverty decreased considerably between 1970 and 2004 (from 15.5 percent to 4.5 percent) and also fell sharply among divorced persons and widow(er)s.6 The never-married elderly also witnessed a reduction in poverty during this period, but it was much less compared with that of the other groups (from 29.2 percent in 1970 to 21.9 percent in 2004). In fact, in the early 1990s, their poverty rate actually rose, which led to a change in relative poverty rates for widowed and divorced persons.

Several factors might explain the concentration of elderly poverty among unmarried groups, especially the never-married. One relates to disparities in lifetime earnings between the married and unmarried (Seigel 1993; Waite 1995; Smock, Manning, and Gupta 1999; Wilmonth and Koso 2002). Another issue pointed out by research is that marriage tends to promote economies of scale in household production such that the married are able to pool resources and share costs of household goods and services, lowering the overall cost of such items if secured individually. Economic models of the family also note the gains of marriage from the specialization of spouses and the division of household labor, whereby each spouse is able to focus on separate skills related to the market or domestic sectors, resulting in more efficiency (Becker 1981; Waite 1995, 493.

The institutional factors associated with marriage may also help explain poverty rates among elderly unmarried groups such as the never-married. One factor to consider is Social Security auxiliary benefits, which represent an important source of retirement income for divorced or widowed women with low lifetime earnings or intermittent employment.7 Pension entitlement may be another aspect, and, not surprisingly, individuals who never married would not have access to spousal pension income during retirement. In this sense, never-married women may experience an economic disadvantage, insofar as women are either less likely to have pensions than men or more likely to have smaller pensions as a result of lower lifetime earnings (Hardy and Shuey 2000). Without access to the pension income of a spouse, never-married women may then have a greater reliance on Social Security retired-worker benefits for income support.

Economic resources. Although the analysis thus far has called attention to the poverty rate among never-married elderly individuals, they are not an economically homogeneous group. Table 2 indicates two distinct segments of the never-married population—one with very little economic resources and one with higher income. Thirty percent of never-married men have annual total money income below $10,000, but 16.5 percent have an annual income that equals or exceeds $40,000 (the corresponding percentages for never-married women are 34.9 percent and 11.3 percent).

Table 2. Selected measures of economic welfare for married couples and nonmarried persons aged 65 or older, by marital group and sex, 2004 (in percent unless otherwise noted)
Measure Married
couples a
Never-married Divorced Widowed
Men Women Men Women Men Women
Distribution by total money income b
Less than $10,000 4.0 30.4 34.9 20.2 33.4 19.1 28.9
$10,000–19,999 15.3 32.9 31.9 30.5 30.8 37.1 46.8
$20,000–39,999 38.4 20.2 21.9 28.9 22.8 26.9 18.9
$40,000 or more 42.2 16.5 11.3 20.3 12.6 16.9 5.5
Median total money income (dollars) 34,900 15,000 14,400 19,979 14,335 18,013 13,003
Percentage with income from specified source
Retirement benefits 93 87 80 89 90 93 93
Social Security c 90 83 75 87 89 89 91
Benefits other than Social Security 51 34 39 38 32 45 32
Other public pensions 18 10 13 11 12 15 11
Railroad retirement 1 1 0 0 0 1 1
Government employee pensions 18 9 13 11 12 13 10
Military 2 3 0 2 1 2 1
Federal 6 2 3 2 3 5 3
State or local 11 5 9 6 9 7 6
Private pensions or annuities 36 25 26 27 20 32 23
Earnings d 37 19 20 26 27 13 10
Income from assets 67 49 46 48 46 50 47
Interest 64 45 43 44 43 46 43
Other income from assets 35 19 21 20 17 21 19
Dividends 29 15 17 14 13 17 14
Rent or royalties 12 10 5 8 5 7 6
Estates or trusts 0 0 0 0 0 0 1
Public assistance
Supplemental Security Income 2 10 9 4 8 3 5
Other 2 10 9 4 8 3 5
Veterans' benefits 5 7 0 6 1 6 3
Annual Social Security benefit
Less than $2,500 1.3 3.9 4.5 1.6 1.8 1.3 1.9
$2,500–4,999 1.8 8.4 8.0 3.6 6.0 4.6 5.0
$5,000–9,999 8.2 38.9 39.7 27.0 52.6 22.6 34.0
$10,000–14,999 18.3 37.4 39.5 49.8 30.9 47.0 45.5
$15,000 or more 70.5 11.3 8.4 18.0 8.8 24.4 13.5
Median annual Social Security benefit (dollars) 18,679 9,799 9,799 11,712 9,199 12,000 10,800
SOURCE: Author's tabulations based on Income of the Population 55 or Older, 2004, Tables 1.1, 1.9, 3.1, 3.7, 5A.2, and 5A.5.
a. The age of a married couple is the age of the husband if he is 65 or older; if the husband is younger than 55 and the wife is aged 65 or older, the age of the married couple is the age of the wife.
b. Total money income refers to the sum of all income received by the married couple or nonmarried person. Income sources can include any source that is regularly received such as wages, salaries, self-employment income, Social Security, Supplemental Security Income, public assistance, interest, dividends, rent royalties, pensions, and so on. Columns may not add to exactly 100 because of rounding.
c. Includes retired-worker benefits, dependents' or survivors' benefits, disability benefits, transitionally insured benefits, and special age-72 benefits.
d. Includes wages, salaries, and self-employment income.

Compared with other unmarried groups, the never-married share some similarities and exhibit some important variations on selected measures of economic welfare. The income distribution of never-married women, for example, is very similar to that of divorced women but less similar when compared with that of widowed women (the never-married have a higher incidence of both low- and high-income persons). Widowed and divorced men are decidedly less likely to have low income (that is, less than $10,000) than never-married men (19.1 percent and 20.2 percent, respectively, compared with 30.4 percent).

In terms of income sources at retirement age, the data again reveal similarities and differences among unmarried subgroups. The vast majority of unmarried persons aged 65 or older (75 percent to 91 percent), both men and women, reported income from a Social Security benefit. Another important source of income for the elderly was a pension or annuity, which around 25 percent of never-married men and women reported receiving. Among unmarried women, the never-married had the highest share (26 percent) with a private pension or annuity. By contrast, among unmarried men, widowers (32 percent) had the greatest proportion receiving pension income and never-married the lowest (25 percent). Also noteworthy is the comparatively high share of never-married men and women receiving Supplemental Security Income (10 percent and 9 percent, respectively).

Breakouts on annual Social Security benefits further illustrate diverse outcomes among the never-married elderly. On this point, the data suggest a relative concentration of never-married men and women with low annual Social Security benefits (less than $4,999). At the same time, the median annual Social Security benefit for never-married women is higher than that for divorced women (widows have the highest median benefit among the three unmarried groups). Never-married men have a lower median benefit than their counterparts in the widowed and divorced groups.

The Health Dimension

The relationship between marital status and health has attracted the increasing attention of researchers (Ross, Mirowsky, and Goldsteen 1990; Lillard and Panis 1996; Murphy, Glaser, and Grundy 1997; Schone and Weinick 1998; Barrett and Lynch 1999; Brown 2000; Simon 2002; Schoenborn 2004; Brown, Roebuck Bulanda, and Lee 2005; Elwert and Christakis 2006). According to research, married persons have, on average, healthier profiles than the unmarried, be they divorced, widowed, or never-married. Research has also found that married individuals live longer than unmarried persons and that never-married men have especially higher mortality rates (Goldman, Korenman, and Weinstein 1995; Lillard and Waite 1995; Rogers, Hummer, and Nam 2000; Waite and Gallagher 2000).8

Recent evidence from the National Health Interview Survey (NHIS) highlights the interplay between marital status and health in old age.9 Table 3, based on NHIS data reported in Schoenborn (2004), shows important differences in health status, limitations, and health-risk behaviors between married and unmarried persons aged 65 or older.10 Across marital groups, widow(er)s were most likely to have reported an activity limitation (46 percent), followed by the never-married (41 percent), divorced (41 percent), and the married (29 percent).11 The never-married had significantly higher proportions reporting restrictions on daily activities (16 percent) and being physically inactive (58 percent) compared with the respective proportions of the married and the national average, but lower proportions than those of the widowed (21.5 percent and 61.4 percent, respectively). The divorced and separated and never-married elderly had similarities on many of the health measures, a notable exception being smoking. The divorced reported a significantly higher rate of current smokers (19 percent) than their widowed and never-married counterparts as well as the married.12

Table 3. Percentage aged 65 or older with selected health characteristics, by marital status, 1999–2002
Characteristic All a Married Never-married Divorced or separated Widowed
Percent Standard
error
Percent Standard
error
Percent Standard
error
Percent Standard
error
Percent Standard
error
Limitation in b
Any activity c 35.7 0.44 28.9 0.53 40.8 1.73 41.3 1.09 45.6 0.61
Work activity d 24.7 0.39 19.9 0.48 29.6 1.64 32.4 1.06 30.7 0.56
ADL or IADL e 12.9 0.27 7.6 0.29 16.0 1.25 14.1 0.79 21.5 0.51
Physically inactive f 53.1 0.48 47.7 0.61 58.3 1.70 54.1 1.12 61.4 0.61
Current smoker g 9.9 0.22 8.3 0.29 10.2 1.01 19.1 0.91 10.0 0.36
Overweight or obese h 57.1 0.38 60.3 0.54 52.2 1.74 57.6 1.11 51.8 0.57
SOURCE: National Health Interview Survey 1999–2002, reported in Schoenborn (2004), Tables 3 and 7.
a. Includes persons "living with a partner" but not married.
b. The Centers for Disease Control and Prevention recommends interpreting this measure of work limitation with caution because many persons aged 65 or older have left the workforce (Schoenborn 2004, 5).
c. Limitation in any activity is based on a series of questions concerning limitations in a person's ability to engage in a variety of activities because of a physical, mental, or emotional problem, including work and school activities, activities because of a physical, mental, or emotional problem, including work and school activities, activities of daily living, instrumental activities of daily living, walking, remembering, or any other unspecified life activities.
d. Limitation in work activity is based on the questions "Does a physical, mental, or emotional problem NOW keep you from working at a job or business?" and for persons not kept from working, "Are you (or any family members) limited in the kind or amount of work they can do because of a physical, mental, or emotional problem?"
e. ADL represents activities of daily living. Results are based on the question "Because of a physical, mental, or emotional problem, does (person) need the help of other persons in personal care needs, such as bathing, dressing, or getting around inside the home?" IADL represents instrumental activities of daily living, such as everyday household chores, doing necessary business, shopping, or getting around for other purposes.
f. "Physically inactive" refers to adults who engaged in no light, moderate, or vigorous leisure-time physical activity, including those who said they were unable to do such activities.
g. "Current smoker" is a person who had ever smoked 100 cigarettes and was smoking as of the date of the interview.
h. "Overweight or obese" is body mass index (BMI) greater than or equal to 25.

Various arguments have been advanced to explain an association between marital status and health. One relates to marriage protection, which views marriage as having a protective economic, social, or psychological effect on health (Waite 1995; Rogers, Hummer, and Nam 2000). Umberson (1987), for example, argues that marriage may promote the kinds of social regulation that buffer negative health inputs, such as smoking and drinking. Using the 1987 National Medical Expenditure Survey, Schone and Weinick (1998) found that married persons of retirement age engage in less risky health behaviors than do their unmarried counterparts (see also Waite [1995, 487], for data on all-age population).13

Alternatively, the association between marital status and health may reflect marriage selection bias. This argument suggests that the protective health effects of marriage are a consequence of the selection into marriage of people who are already healthy—that is, healthy people are more likely than less healthy people to marry and stay married (Goldman 1993). Although there is still much debate on the causal direction and processes mediating the effects of marital status on health (Goldman, Korenman, and Weinstein 1995), the marital status composition of the elderly is likely to influence the health and care costs of the aged.

Never-Married Retirees in the Future

To assess trends among the never-married elderly in the near future, this section turns to projection data. Analysis is based on SSA's Modeling Income in the Near Term (MINT) model. Developed by SSA's Office of Research, Evaluation, and Statistics with assistance from the Brookings Institution, the RAND Corporation, and the Urban Institute, MINT is a powerful tool for analyzing the expected economic and demographic characteristics of future retirement populations.

For birth cohorts from 1926 to 1972, MINT uses respondents to the 1990–1993 and 1996 panels of the Survey of Income and Program Participation (SIPP).14 Observed data and statistically estimated parameters are then used to project wealth, income (for example, total family income and Social Security benefits), and demographic outcomes (for example, marital status and mortality) into the future.

Recently, MINT has been extended beyond its original "near term" structure to include later birth cohorts, namely, the 1973–2017 cohorts. Because these cohorts would not have reached their mid-20s by the time the SIPP panels were fielded, the Urban Institute developed special procedures to project their future circumstances. In essence, MINT respondents from the late baby-boom cohorts (1960–1964) are selected to represent these later cohorts. The selected respondents are designed to match the later cohorts on observed or projected characteristics.15 A detailed discussion of the MINT-extended design is beyond the scope of this article (for a fuller discussion, see Butrica and others 2001; Toder and others 2002; Smith, Cashin, and Favreault 2005, Chapter 5), but it is important to keep in mind the greater uncertainty associated with the projections for the later cohorts. Results presented here for retirees in 2020–2030 are more reliable than those for retirees in 2040–2060.

The categories of marriage, divorce, and widowhood in MINT are based on survey responses in the SIPP panels and statistical models developed to predict such events. Gender-specific, continuous time hazard models predict marriage formation, divorce, and remarriage (Panis and Lillard 1999). Explanatory variables include age, education, race and ethnicity, years unmarried, whether widowed, and the calendar year after 1980 to control for time trends in marriage. Statistical models and other techniques are used for mortality projections that underlie, among other things, projections of widowhood. Characteristics of current, former, and future spouses are also estimated by MINT. Individuals married at the time of the SIPP panels and projected to remain married are matched to their spouse in the survey. Characteristics of former and future spouses are imputed by MINT and then statistically assigned to a MINT observation with similar characteristics using the "nearest neighbor" method. MINT projections of income sources in retirement involve careful modeling of their determinants, most notably earnings over the lifetime. Toder and others (2002, II-10) discuss MINT's projections of earnings over the life cycle and note the ability of those projections to capture important labor market changes, such as the increased participation of women in the labor force.

Projections in this article are restricted to persons aged 62 or older for 2020, 2030, 2040, 2050, and 2060. Analysis is organized around three particular issues:

Projected Marital Composition and Demographics

Consistent with demographic trends showing increases in never-married persons among younger cohorts, MINT projects a shift in the marital status composition of future retirees. This change is marked by an increasing proportion of unmarried persons aged 62 or older, namely, the never-married: 4 percent in 2020, 7 percent in 2030, and 10 percent in 2060 (Chart 3).16 Put another way, in future years a smaller proportion of retirees will be in a currently married status compared with the proportion today. The share of widowed retirees is projected to increase but more slowly than that of the never-married. Meanwhile, the proportion of divorced persons is expected to remain fairly constant, reflecting the leveling out of divorce rates over time. This picture is consistent with other models that project the marital composition of the retirement-age population in future years (see, for example, Favreault and Smith 2004).

Chart 3.
Projected distribution of retirement-age persons (aged 62 or older), by marital status, 2020–2060
Line chart with tabular version below.
Show as table
Table equivalent for Chart 3. Projected distribution of retirement-age persons (aged 62 or older), by marital status, 2020–2060
Year Never-married Divorced Widowed Married
2020 5.20 15.60 19.50 59.80
2030 6.60 16.60 21.10 55.70
2040 8.10 16.40 23.40 52.10
2050 9.40 16.10 24.80 49.70
2060 10.40 15.90 23.60 49.90
 
SOURCE: Author's calculations using data from Modeling Income in the Near Term (MINT).

Demographic projections offer a sharper picture of future never-married retirees (Table 4). With respect to age structure, never-married retirees are more likely to be in the youngest age group (62–69) than in the overall retirement-age population (aged 62 or older). The difference is greatest in 2020, when 59.4 percent of never-married retirees are projected to be in the youngest age group compared with 46.1 percent from the overall group. Over time, however, the difference narrows. Also notable, MINT estimates that women will account for a slightly larger proportion of future never-married retirees than of the retirement-age population ("All" in Table 4)—60 percent and 57 percent, respectively—in 2020. This pattern is expected to hold through 2060.

Table 4. Projected demographic characteristics of the never-married and the total retirement-age population aged 62 or older, 2020–2060 (in percent)
Characteristic 2020 2030 2040 2050 2060
Never-
married
All Never-
married
All Never-
married
All Never-
married
All Never-
married
All
Age
62–69 59.4 46.1 49.0 37.4 41.3 31.5 38.3 32.6 37.6 33.9
70–79 29.2 35.5 36.5 38.5 37.1 35.8 33.5 31.2 32.5 32.6
80–89 9.8 14.9 11.9 18.7 17.5 23.8 20.3 23.5 19.3 20.2
90 or older 1.6 3.4 2.7 5.4 4.2 8.9 8.0 12.7 10.6 13.3
Sex
Male 39.8 43.1 40.4 41.5 38.4 40.4 35.5 40.0 36.2 40.4
Female 60.2 56.9 59.6 58.5 61.6 59.6 64.5 60.0 63.8 59.6
Race and ethnicity a
Non-Hispanic
White 66.9 77.6 61.4 74.1 57.6 69.7 52.9 66.9 48.7 63.4
Black 21.3 8.8 23.0 9.2 23.4 9.8 26.7 11.0 27.1 11.5
Hispanic 9.0 7.9 11.1 10.0 13.5 13.3 13.9 15.3 16.4 17.5
Asian-American 2.8 5.6 4.3 6.7 5.5 7.3 6.5 6.9 7.9 7.6
Education
Less than high school 13.8 12.0 10.8 10.4 9.4 9.7 9.8 9.4 8.3 9.0
High school 51.0 60.0 54.6 60.3 54.9 59.3 56.5 57.6 57.7 56.2
College graduate 35.2 28.0 34.6 29.4 35.7 31.0 33.7 33.0 34.0 34.8
SOURCE: Author's calculations using data from Modeling Income in the Near Term (MINT).
NOTE: The never-married samples are relatively small compared with the "All" category (those aged 62 or older), but in no case are they below 2,734 observations in MINT.
a. Race and ethnicity are based on self-identification for which respondents could select only one race category (rather than multiple categories).

MINT also estimates future marital status by race and ethnicity. Results show that while the majority of future never-married retirees are expected to be white, a sizable percentage are expected to be black—21 percent of the never-married population aged 62 or older in 2020 but only 9 percent of the entire population of the same age group. By 2060, the share of blacks among the never-married is expected to increase to 27 percent but to only 11.5 percent among the overall retirement-age population.17 Additionally, MINT projects a rise in the share of Hispanics among the never-married and the retirement-age population as a whole because of the dramatic growth in the Hispanic population in the United States.

MINT estimates of educational attainment are also of interest. Overall, MINT projects an increasingly well-educated retirement-age population (for example, baby boomers are more likely to be college graduates and less likely to be high school dropouts than are current retirees). The diverse nature of the never-married population is evident in terms of their educational attainment as future retirees. One segment will be college educated—according to MINT, 35 percent of the never-married population aged 62 or older will have a college degree in 2020 compared with 28 percent among the general retirement-age population. This college-educated segment will be in contrast, however, to a segment of never-married high school dropouts (14 percent of the never-married in 2020 compared with 12 percent overall). Since education tends to be highly correlated with lifetime earnings, never-married retirees without a high school degree will probably represent some of the persons most likely to experience economic difficulties in retirement in the near future (the proportions of the never-married and general retiree populations without a high school degree are projected to decline from 2020 to 2060).

Projected Economic Well-Being

This section examines the projected economic situation of never-married retirees using several measures: poverty rate, income distribution, and welfare ratio.

Poverty Rates. A decline in elderly poverty rates between 2020 and 2060 can be observed in Table 5, for the overall retirement-age population and for the marital groups. A major contribution to the decrease in elderly poverty rates in future years is the assumption that wage growth will exceed price growth, which results in greater retirement income for future retirees because Social Security benefits are indexed to wages (and because a constant saving rate produces higher real savings). This result is consistent with other research suggesting that baby boomers, as well as younger cohorts, will experience greater absolute economic security in retirement than current retirees.

Table 5. Projected poverty rate of retirees aged 62 or older, by marital status, 2020–2060 (in percent)
Marital status 2020 2030 2040 2050 2060
All 5.1 4.5 4.0 2.7 1.8
Married 3.1 2.9 2.4 1.4 1.0
Never-married 13.4 11.5 10.3 8.5 6.5
Divorced 7.4 5.6 5.0 3.2 2.4
Widowed 7.1 5.6 4.4 2.6 1.1
SOURCE: Author's calculations using data from Modeling Income in the Near Term (MINT), assuming current law and scheduled Social Security benefits.
NOTE: The poverty threshold used for each person is based on family size and elderly status and is indexed to price growth. Household income used to determine poverty status includes earnings, private pension income, income received from annuitizing assets, Social Security benefits, Supplemental Security Income benefits, and income from any nonspouse coresident.

However, the magnitude of elderly poverty is expected to continue to vary across marital groups in future years. MINT projects that the never-married will have the highest poverty rate of all other marital groups between 2020 and 2060. By 2040, the poverty rate among the never-married (10.3 percent) is projected to be about three-quarters of their rate (13.4 percent) in 2020, which approximates the percentage decline experienced by all persons aged 62 or older. Between 2040 and 2060, however, elderly poverty declines less quickly for the never-married (the rate for the never-married in 2060 is 63 percent of their 2040 rate, compared with 45 percent for all persons). This result is surprising in that the never-married, over time, are becoming a less "select" group (that is, a greater percentage of the overall population) and might be expected to have trends that mimic the broader population. Because projections for years beyond 2040 are less reliable, these latter results should be viewed with caution.

Income Distribution and Welfare Ratios. Measures of income distribution and welfare ratios offer further insight (Table 6). With respect to income composition, MINT projects that a relatively high share of never-married retirees will be in the lowest income quintile and that the share will increase (from around 31 percent in 2020 to 35 percent in 2060). At the same time, a substantial segment of the never-married is expected to reach the highest income distribution cutoffs (37 percent will reach the two highest income quintiles in 2020). This high-income segment of the never-married is projected, however, to decrease slightly over time.

As a final point of analysis, a welfare ratio measure was calculated for each marital group. A welfare ratio expresses the ratio of family income to household needs (set to the appropriate U.S. poverty thresholds for the elderly aged 65 or older). It is a good gauge of economic well-being because it adjusts for household size, accounts for the different needs of families, and uses the entire income distribution of the population of interest. Moreover, unlike per capita income measures, a welfare ratio uses poverty thresholds, which helps account for economies of scale by assuming that those who are married need 1.26 times more income (rather than 2 times) to live equally as well as a nonmarried person.

The data in Table 6 show that the projected median welfare ratio of the never-married will be below the national median and below that of married persons as well as that of the divorced and widowed from 2020 through 2060. For example, the median family income for never-married persons aged 62 or older in 2020 is estimated to be 3.3 times the poverty threshold. By contrast, the median family income will be 5.9 times the poverty threshold for married individuals and 4.9 times the threshold for the overall median. Calculating the welfare ratio for the never-married as a ratio of other marital groups suggests that the never-married will have, on average, proportionally less income in retirement than other marital groups. For example, expressed in relation to the never-married, the projected welfare ratio for the divorced is 1.12 in 2020—that is, 12 percent higher than that of the never-married. Likewise, MINT projects the median welfare ratio of widows to be 9 percent higher than that of the never-married in 2020 and the married a striking 79 percent higher than that of the never-married. This pattern is predicted to hold and in some cases increase through 2060, assuming that scheduled benefits are in place according to current law.

Table 6. Projected measures of economic welfare for retirees aged 62 or older, by income quintile and marital status, 2020–2060
2020 2030 2040 2050 2060
Percentage distribution of the never-married aged 62 or older, by income quintile a
Lowest 30.8 31.2 31.1 34.0 35.4
Second lowest 16.1 15.7 19.6 19.6 19.7
Middle 16.5 16.6 15.3 15.0 14.8
Second highest 17.4 18.6 17.1 17.6 15.8
Highest 19.4 17.9 17.1 14.9 14.4
Median welfare ratio of all aged 62 or older, by marital status b
All 4.9 5.2 5.1 5.8 6.5
Married 5.9 6.2 5.6 6.6 7.3
Never-married 3.3 3.7 3.7 3.9 4.2
Divorced 3.7 4.0 4.2 4.9 5.5
Widowed 3.6 4.2 4.5 5.2 6.2
Ratio of median welfare of all aged 62 or older who are divorced, widowed, or married to that of the never-married c
All 1.48 1.41 1.38 1.49 1.55
Married 1.79 1.68 1.51 1.69 1.74
Never-married . . . . . . . . . . . . . . .
Divorced 1.12 1.08 1.14 1.26 1.31
Widowed 1.09 1.14 1.22 1.33 1.48
SOURCE: Author's calculations using data from Modeling Income in the Near Term, assuming current law and scheduled Social Security benefits.
NOTE: . . . = not applicable.
a. The income quintile cutoffs are based on all persons aged 62 or older in MINT for each analyzed year.
b. The welfare ratio refers to the ratio of family income to U.S. poverty thresholds for the elderly. Family income used to compute a welfare ratio includes earnings, pension income, income received from annuitizing assets, Social Security benefits, Supplemental Security Income benefits, plus income from any nonspouse coresident. Imputed rent is not included in the family income measure used to determine welfare ratios.
c. For example, the median welfare ratio of divorced to that of the never-married is 1.12.

Discussion

It is important that policymakers have wide-ranging information about groups with a greater likelihood of economic insecurity at retirement. This article sheds light on one group of concern, the never-married. An important finding is that the vast majority of elderly Americans, regardless of marital status, live above the poverty line. In the future, the share of elderly Americans in poverty is projected to continue to decline, in part because of the increasing real value of Social Security benefits. This article demonstrates, though, that never-married elderly Americans are relatively more likely to experience economic and health difficulties, including poverty, than other marital groups.

This article also offers insight into the circumstances of future never-married retirees. MINT projects that the proportion of retirees who have never-married will rise between 2020 and 2060, suggesting that more persons retiring in the future will receive only a retired-worker benefit. Another finding is that a sizable percentage of future never-married retirees will be black, largely because of differences in the marriage rates of white and black Americans. A further issue to consider is the projected growth of the Hispanic population over the next 40 years, among the never-married and among the elderly population as a whole.

According to MINT, elderly poverty is projected to decline across all marital subgroups; however, the never-married will continue to experience a higher incidence of economic insecurity at retirement age compared with other marital groups. The never-married are expected to have both the highest poverty rate and the lowest median welfare ratio among marital groups between 2020 and 2060.18 Moreover, the share of never-married elderly located at the lower end of the income distribution is projected to be relatively sizable in 2020 and on the increase. In other words, although widows receive a great deal of attention in the Social Security debate because of their incidence of poverty, the projections for the never-married, on average, show even higher poverty rates. Such outcomes call attention to the circumstances of never-married retirees and more broadly to marital trends in relation to the Social Security program.

Heterogeneity among never-married retirees must also be recognized. Varied circumstances among never-married retirees can be observed in the distribution of income as well as in educational attainment. Around 17 percent of never-married men and 11 percent of never-married women (65 or older) reported total income of $40,000 or more in 2004, almost half had income from assets, and around 25 percent received some income from private pensions or annuities. The future never-married elderly are projected to have a higher than average share of persons with a college degree in 2020 but also a slightly higher percentage of persons without a high school diploma. Thus, the never-married may be expected to exhibit diverse economic outcomes.

Given the looming sociodemographic changes in the retiree population as the first wave of the baby-boom cohort begins to reach retirement age, an important area of study is the relationship between family structures and retirement outcomes. Although considerable research has documented the importance of marital status on adult well-being, work specifically focused on the elderly population is more limited. This analysis demonstrates that research is needed on different categories of unmarried retirees. Finally, multivariate analysis that estimates the net effect of marital status on economic and health well-being in old age, while controlling for factors such as education level, race and ethnicity, and age, would be useful. Such focal points would help explore the retirement needs of all elderly Americans.

Notes

1. The increasing rate of unmarried adults has been associated with various factors, including the ratio of women's earnings to men's, no-fault divorce legislation, and social and cultural factors, such as the increase in cohabitation and divorce.

2. Social Security auxiliary benefits are based on a retiree's current marital status or marital history. Married individuals can qualify for a spousal benefit at retirement. A spousal benefit can equal 50 percent of their partner's primary insurance amount (PIA), based on the partner's earnings history. Persons divorced from a retired worker or disabled worker may also qualify for a spousal benefit, provided that they were married for at least 10 years. Survivor benefits, which can equal 100 percent of the worker's PIA, are based on marital history (in addition to a worker's earnings history). Reduced survivor benefits are payable to widows as early as at age 60, provided that their deceased spouse worked in employment covered by Social Security and that they have not remarried before the age of 60 (SSA 2006a).

3. Most of the literature is based on U.S. data. However, recent research shows increasing interest in international comparisons. Steinsultz (2006), for example, compares the level of wealth of never-married women across Canada, Germany, Sweden, and the United States using the Luxembourg Income Study (LIS).

4. See the appendix in SSA's (2006b) Income of the Population 55 or Older, 2004, for methods on computing standard errors and testing for differences between two sample estimates.

5. Poverty rates also vary by age group (65–74, 75 or older), education, and race and ethnicity.

6. The poverty rate of widow(er)s dropped from around 35 percent in 1970 (the marital group with the highest elderly poverty rate at this time) to 14.5 percent in 2004, and that of divorced persons from 32.5 percent to 17 percent. Changes have been made in the survey over time, which are discussed in some detail in U.S. Census Bureau, Current Population Reports, Series P60 (various years).

7. More than 31 percent of women beneficiaries aged 62 or older in 2004 were entitled solely on the basis of a current, former, or deceased spouse (SSA 2006a, Table 5.A14). Another 28 percent were dually entitled to spouse or survivor benefits. Dual entitlement occurs when a person is also entitled to a worker benefit (the full worker benefit and a partial spouse or survivor benefit are paid in these cases).

8. Although the marriage mortality "advantage" holds for both men and women, it is greater for men (Lillard and Panis 1996).

9. The NHIS represents a major data source of health behavior and conditions of the noninstitutionalized U.S. civilian population. The survey is conducted by the National Center for Health Statistics of the Centers for Disease Control and Prevention and reflects a nationally representative sample.

10. The Centers for Disease Control and Prevention recommends interpreting the work limitation measure in Table 3 with caution because many in the 65 or older group have left the workforce (Schoenborn 2004, 5).

11. The never-married may be worse off than widow(er)s when controlling for age. The percentage difference between the divorced and never-married is not statistically significant (at the 0.05 level).

12. Differences in the health behaviors across marital groups can be even larger among younger and middle-aged groups (Schoenborn 2004, 4).

13. Rogers, Hummer, and Nam (2000) give evidence that divorced persons have the highest occurrence of morbidity and mortality and that widow(er)s' health and lifestyle behaviors are closer to those of married persons.

14. SIPP is a survey of the U.S. civilian noninstitutionalized population. Because the survey design is a continuous national panel, SIPP supports both longitudinal and cross-sectional data analysis. Interviews are conducted every 4 months for 28 to 36 months. The survey provides robust information on income and wealth, labor force participation, participation in government programs, marital histories, and a host of other socioeconomic and demographic variables that allow measurement of the future costs and effectiveness of existing government programs.

15. Specifically, the 1960–1964 MINT respondents serve as potential "donors" in statistical matches to respondents in the 2003 Current Population Survey (CPS) or, for the 1984–2017 cohorts, to simulated individuals in Census Bureau projections. Matching variables from the CPS include age, sex, race and ethnicity, education (less than high school, high school graduate, and college graduate), age-specific earnings, age-specific marital status, and foreign-born status. Matching variables for Census projections include only sex, race and ethnicity, and foreign-born status.

16. These figures correspond with population projections from SSA's Office of the Chief Actuary (see Bell 1997, Table 21C).

17. Along this line, Goldstein and Kenny (2001) approximated that among women born between 1960 and 1964, 93 percent of whites will ever marry compared with 64 percent of blacks (see also Harrington Meyer, Wolf, and Himes 2004).

18. Changes in patterns of economic growth, labor force participation, and real wage growth could alter these figures.

References

Barrett, Anne E., and Scott M. Lynch. 1999. Caregiving networks of elderly persons: Variation by marital status. Gerontologist 39(6): 695–704.

Becker, Gary S. 1981. A treatise on the family. Cambridge, MA: Harvard University Press.

Bell, Felicitie C. 1997. Social Security area population projections. Actuarial Study No. 112. Baltimore: Social Security Administration, Office of the Chief Actuary.

Brown, Susan. 2000. The effects of union type on psychological well-being: Depression among cohabitors versus marrieds. Journal of Health and Social Behavior 41(3): 241–255.

Brown, Susan L., Jennifer Roebuck Bulanda, and Gary R. Lee. 2005. The significance of nonmarital cohabitation: Marital status and mental health benefits among middle-aged and older adults. Journal of Gerontology Series B, Psychological and Social Sciences 60(1): S21–S29.

Butrica, Barbara A., and Howard M. Iams. 2000. Divorced women at retirement: Projections of economic well-being in the near future. Social Security Bulletin 63(3): 3–12.

Butrica, Barbara A., Howard M. Iams, James Moore, and Mikki Waid. 2001. Methods in Modeling Income in the Near Term (MINT). ORES Working Paper No. 91. Washington, DC: Social Security Administration, Office of Policy.

Cherlin, Andrew J. 1992. Marriage, divorce, and remarriage, rev edition Cambridge, MA: Harvard University Press.

Easterlin, Richard A., Christine M. Schaeffer, and Diane J. Macunovich. 1993. Will the baby boomers be less well off than their parents? Income, wealth, and family circumstances over the life cycle in the United States. Population and Development Review 19(3): 497–522.

Elwert, Felix, and Nicholas A. Christakis. 2006. Widowhood and race. American Sociological Review 71(1): 16–41.

Favreault, Melissa, and Karen Smith. 2004. A primer on the Dynamic Simulation of Income Model (DYNASIM3). Discussion Papers 02-04, Urban Institute, Washington, DC.

Goldman, Noreen. 1993. Marriage selection and mortality patterns: Inferences and fallacies. Demography 30(2): 189–198.

Goldman, Noreen, Sanders Korenman, and Rachel Weinstein. 1995. Marital status and health among the elderly. Social Science and Medicine 40(12): 1717–1730.

Goldstein, Joshua R. 1999. The leveling of divorce in the United States. Demography 36(3): 409–414.

Goldstein, Joshua R., and Catherine T. Kenney. 2001. Marriage delayed or marriage forgone? New cohort forecasts of first marriage for U.S. women. American Sociological Review 66 (August): 506–519.

Gustman, Alan, and Thomas Steinmeier. 2000. Retirement in dual-career families: A structural model. Journal of Labor Economics 18(3): 505–545.

Hardy, Melissa A., and Kim Shuey. 2000. Pension decisions in a changing economy: Gender, structure, and choice. Journal of Gerontology Series B, Psychological and Social Sciences 55(5): S271–S277.

Harrington Meyer, Madonna, Douglas A. Wolf, and Christine L. Himes. 2004. Linking benefits to marital status: Race and diminishing access to Social Security spouse and widow benefits in the U.S. CRR Working Paper 2004-05, Center for Retirement Research at Boston College, Chestnut Hill, MA. Available at www.bc.edu/centers/crr/papers/wp_2004-05.pdf.

———. 2006. Declining eligibility for Social Security spouse and widow benefits in the United States? Research on Aging 28(2): 240–260.

He, Wan, Manisha Sengupta, Victoria A. Velkoff, and Kimberly A. DeBarros. 2005. 65+ in the United States: 2005. Current Population Reports, P23-209. Washington, DC: U.S. Census Bureau.

Keith, Pat. 2003. Resources, family ties, and well-being of never-married men and women. Journal of Gerontology Series B, Psychological and Social Sciences 42(2): 51–75.

Kreider, Rose M. 2005. Number, timing, and duration of marriages and divorces: 2001. Current Population Reports, P70-97. Washington, DC: U.S. Census Bureau.

Kreider, Rose M., and Tavia Simmons. 2003. Marital status: 2000. Census 2000 Brief, C2KBR-30, U.S. Census Bureau, Washington, DC.

Lillard, Lee A., and Constantijn W.A. Panis. 1996. Marital status and mortality: The role of health. Demography 33(3): 313–327.

Lillard, Lee A., and Linda J. Waite. 1995. 'Til death do us part: Marital disruption and mortality. American Journal of Sociology 100(5): 1131–1156.

Martin, Patricia, and David Weaver. 2005. Social Security: A program and policy history. Social Security Bulletin 66(1): 1–15.

Morgan, Leslie A. 1992. Marital status and retirement plans: Do widowhood and divorce make a difference? In Families and retirement, editors Maximiliane Szinovacz, David J. Ekerdt, and Barbara H. Vinick, 114–126. Newbury Park, CA: Sage Publications.

Murphy, Mike, Karen Glaser, and Emily Grundy. 1997. Marital status and long-term illness in Great Britain. Journal of Marriage and the Family 59(1): 156–164.

Norton, Arthur J., and Louisa F. Miller. 1992. Marriage, divorce, and remarriage in the 1990s. Current Population Reports, Special Studies P23-180. Washington, DC: U.S. Census Bureau.

Panis, Constantijn, and Lee Lillard. 1999. Near-term model development part II. Final Report, SSA Contract No. 600-96-27335. Santa Monica, CA: RAND.

Pienta, Amy Mehraban, and Mark D. Hayward. 2002. Who expects to continue working after age 62? The retirement plans of couples. Journal of Gerontology Series B, Psychological and Social Sciences 57(4): S199–S208.

Rogers, Richard G., Robert A. Hummer, and Charles B. Nam. 2000. Living and dying in the USA. New York: Academic Press.

Ross, Catherine E., John Mirowsky, and Karen Goldsteen. 1990. The impact of the family on health: The decade in review. Journal of Marriage and the Family 52(4): 1059–1078.

Ruggles, Steven. 1997. The rise of divorce and separation in the United States, 1880–1990. Demography 34(4): 455–466.

Schoenborn, Charlotte A. 2004. Marital status and health: United States, 1999–2002. Advance Data from Vital and Health Statistics No. 351. National Center for Health Statistics, Hyattsville, MD.

Schone, Barbara S., and Robin M. Weinick. 1998. Health-related behaviors and the benefits of marriage for elderly persons. Gerontologist 38(5): 618–627.

Seigel, Jacob S. 1993. A generation of change: A profile of America's older population. New York: Russell Sage Foundation.

Simon, Robin W. 2002. Revisiting the relationships among gender, marital status, and mental health. American Journal of Sociology 107(4): 1065–1096.

Smith, Karen E., David B. Cashin, and Melissa M. Favreault. 2005. Modeling Income in the Near Term 4. Final Report, SSA Contract No. 0600-01-60123. Washington, DC: Urban Institute.

Smock, Pamela J., Wendy D. Manning, and Sanjiv Gupta. 1999. The effect of marriage and divorce on women's economic well-being. American Sociological Review 64(6): 794–812.

[SSA] Social Security Administration. 2006a. Annual Statistical Supplement to the Social Security Bulletin, 2005. Washington, DC: Office of Policy.

_________. 2006b. Income of the Population 55 or Older, 2004. Washington, DC: Office of Policy.

Steinsultz, Linda. 2006. Inequality of wealth for never married women in Canada, Germany, Sweden, and the United States. Luxembourg Income Study Working Paper No. 437. Available at http://www.lisproject.org/publications/liswps/437.pdf, August 20, 2007.

Toder, Eric, Lawrence Thompson, Melissa Favreault, Richard Johnson, Kevin Pevese, Caroline Ratcliffe, Karen Smith, Cori Uccello, Timothy Waidmann, Jillian Berk, Romina Woldermariam, Gary Burtless, Claudia Sahm, and Douglas Wolf. 2002. Modeling Income in the Near Term—Revised projections of retirement income through 2020 for the 1931–1960 birth cohorts. Final Report, SSA Contract No. 600-96-27332. Washington, DC: Urban Institute.

Umberson, Deborah. 1987. Family status and health behaviors: Social control as a dimension of social integration. Journal of Health and Social Behavior 28(3): 306–319.

Waite, Linda J. 1995. Does marriage matter? Demography 32(4): 483–507.

Waite, Linda J., and Maggie Gallagher. 2000. The case for marriage: Why married people are happier, healthier, and better off financially. New York: Doubleday.

Weaver, David A. 1997. The economic well-being of Social Security beneficiaries, with an emphasis on divorced beneficiaries. Social Security Bulletin 60(4): 3–17.

———. 2002. The widow(er)'s limit provision of Social Security. Social Security Bulletin 64(1): 1–15.

Wilmonth, Janet, and Gregor Koso. 2002. Does marital history matter? Marital status and wealth outcomes among preretirement adults. Journal of Marriage and Family 64(1): 254–269.