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. 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. 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. 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. 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. 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. Meanwhile, the share of never-married men and women among the 45—64 age group was roughly double 8. 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. 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. Census Bureau 2000 data, presented in Kreider and Simmons 2003 , Table 1. Includes spouses present and absent. 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. 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. Elderly poverty rates differ not only across marital subgroups but also by sex Chart 1. 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. Tests indicate
