Social Science Research 43 (2014) 60–73

Contents lists available at ScienceDirect

Social Science Research journal homepage: www.elsevier.com/locate/ssresearch

Same-sex cohabiting elders versus different-sex cohabiting and married elders: Effects of relationship status and sex of partner on economic and health outcomes Amanda K. Baumle ⇑ The University of Houston, United States

a r t i c l e

i n f o

Article history: Received 17 December 2012 Revised 3 July 2013 Accepted 1 September 2013 Available online 24 September 2013 Keywords: Gay and Lesbian Aging Marriage Demography

a b s t r a c t In this article, I use pooled data from the 2008–2010 American Community Surveys to examine outcomes for different-sex married, different-sex cohabiting, and same-sex cohabiting elders across several key economic and health indicators, as well as other demographic characteristics. The findings suggest that elders in same-sex cohabiting partnerships differ from those in different-sex marriages and different-sex cohabiting relationships in terms of both financial and health outcomes, and that women in samesex cohabiting partnerships fare worse than men or women in other couple types. The results indicate that financial implications related to the sex of one’s partner might be more predictive of economic and health outcomes in old age, rather than solely access to legal marriage. Nonetheless, findings suggest that individuals in same-sex cohabiting partnerships might experience worse outcomes in old age as a result of cumulative effects across the life course from both the sex of their partner (in the case of female couples) as well as their lack of access to benefits associated with marriage. Accordingly, these findings demonstrate that persons in same-sex cohabiting partnerships require unique policy considerations to address health and economic concerns in old age. Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction In the United States, an aging population continues to raise policy concerns regarding how to best provide financial, health, and social support for varying segments of the elderly. Research indicates that individuals’ needs in old age vary across demographic characteristics such as gender, class, race and ethnicity, education, relationship status, and related characteristics (Uhlenberg, 2009; Lauderdale, 2001; South and Tolnay, 1992). Burgeoning research in the area of sexuality studies also suggests that gay and lesbian individuals experience aging differentially, requiring unique policy considerations (IOM, 2011; Fredriksen-Goldsen and Muraco, 2010; MetLife, 2010; Scherrer, 2009). Most of this research, however, has focused on the experiences of gay and lesbian individuals rather than couples. Accordingly, little is known regarding whether or how outcomes differ for elderly individuals cohabiting in same-sex partnerships as compared with those living in different-sex cohabiting or marital relationships. Although social science research strongly supports the evidence of several ‘‘marriage benefits’’ that accrue from heterosexual marriage (see e.g. Waite, 2005; Seltzer, 2004; Brown, 2000; Horwitz and White, 1998; Brines and Joyner, 1999), it is unclear as to the manner in which lack of access to legal marriage affects outcomes for same-sex partners (see e.g. Badgett, ⇑ Address: Department of Sociology, University of Houston, 450 Philip Hoffman Hall, Houston, TX 77204, United States. E-mail address: [email protected] 0049-089X/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ssresearch.2013.09.003

A.K. Baumle / Social Science Research 43 (2014) 60–73

61

2009; Balsam et al., 2008; Rothblum, 2005; Solomon et al., 2004). Given that most gay men and lesbians have not had the option of legal marriage to a same-sex partner, it might be expected that elders living in same-sex partnerships would demonstrate worse financial and health outcomes than those in different-sex marriages. This would particularly be the case for those who have lived all or most of their life outside of the marital institution, since the rewards from marriage can be expected to accrue over the life course. In this article, I use pooled data from the 2008–2010 American Community Surveys to analyze outcomes for different-sex married, different-sex cohabiting, and same-sex cohabiting elders across several key economic and health indicators. Results highlight the potential effect of both sex and institutional exclusion on outcomes for elders in same-sex partnerships.

2. The benefits of marriage 2.1. Empirical findings and theoretical explanations Social science research has supported the existence of several benefits derived from marriage when compared to cohabitation (Waite, 2005; Seltzer, 2004; Brown, 2000; Horwitz and White, 1998; Waite and Gallagher, 2000).1 Prior research indicates that cohabiting couples vary from married couples in terms of income, participation in the labor force, commingling of assets, presence of children, relationship duration, and a variety of other factors (see e.g. Waite, 2005; Brown, 2000; Smock, 2000; Waite and Gallagher, 2000; Brines and Joyner, 1999; Horwitz and White, 1998; Waite, 2000). Financially, married persons amass more assets than do cohabiting persons (Lupton and Smith, 2003), and appear to benefit due to a specialization of economies of scale (Waite, 2005). Further, it has been argued that this economic advantage, coupled with the social support provided by a spouse, facilitates better physical and emotional health for married persons (Waite, 2005; Mirowsky and Ross, 1999). In contrast, cohabiting individuals are less likely ‘‘to commingle financial resources (Brines and Joyner, 1999);. . .less likely to share leisure time and a social life (Clarkberg et al., 1995); less likely to have children (Bachrach 1987); and less likely to remain together (Smock 2000)’’ (Waite, 2005: 91). Although the gap between married and cohabiting persons varies across countries, research suggests that many of the marital advantages exist across a variety of cultural and geographic contexts (Soons and Kalmijn, 2009; Seltzer, 2004). These benefits are particularly salient in the United States, where most individuals ultimately marry and cohabitation remains primarily a ‘‘practice run’’ for marriage (Seltzer, 2004: Bumpass and Lu, 2000). The two principal theories proffered for the ‘‘marriage benefit’’ concern marital selectivity and protection/benefits derived from the marital institution (Soons and Kalmijn, 2009; Waite, 2005). The marital selectivity theory suggests that those who are healthier and have greater socioeconomic status are advantaged in the marriage market, resulting in married persons performing better on many socioeconomic and health indicators (Waite, 2005). This would indicate that better outcomes for married individuals are not a consequence of the union itself, but are a function of those who are likely to enter into marriage. The more institutional theory suggests that marriage provides protection via perceived permanence, the sharing of social and economic resources as a type of ‘‘insurance,’’ and social attachment to the spouse, the spouse’s network, and other social institutions (Soons and Kalmijn, 2009; Waite, 2000). According to this argument, married persons’ behaviors and resources change as a consequence of marriage, resulting in economic and health well-being compared to non-married persons (Waite, 2005). Cohabiting individuals, on the other hand, are less likely to experience these benefits due to a lack of permanence and a corresponding hesitancy to pool assets. It is this second, institutional theory that is particularly pertinent when examining differentials across same-sex cohabiting elders and different-sex married elders. If the institution of marriage provides tangible benefits, then same-sex couples who are unable to access this institution would be expected to have worse outcomes than heterosexual couples who marry. In the next section, I explore whether we should expect elders in same-sex cohabiting partnerships to look more like cohabiting heterosexual couples than married heterosexual couples, in terms of their demographic, economic, and health characteristics.

2.2. Marriage and cohabitation for same-sex elderly couples According to the institutional theory, elders living in same-sex partnerships would be expected to evidence the repercussions of exclusion from the marital institution – both in the present, as well as due to the cumulative lifetime effects for those who have spent much or all of their lives outside of heterosexual marriage. In this respect, marital benefits likely operate in much the same way as other socioeconomic disparities. For example, elders with less education have incurred the cumulative disadvantage of a lifetime of differential access to benefits that would promote health and welfare (Lauderdale, 2001). If marriage produces economic, social, and health benefits, it might be expected that elder same-sex partners who have spent much or all of their lives outside of heterosexual marriage would reflect the cumulative disadvantage of their exclusion from 1 Notably, some studies have not found differences in outcomes for cohabiting and married persons (see e.g. Musick and Bumpass, 2006; Brown, 2000; Brown and Booth, 1996).

62

A.K. Baumle / Social Science Research 43 (2014) 60–73

the institution. Further, these disadvantages might be more strongly experienced by this cohort of elderly due to the increased likelihood of living a closeted life or experiencing frequent and direct discrimination (Herdt and de Vries, 2004). Given that access to legal marriage is relatively recent for same-sex couples, the literature examining the potential benefits accruing from marriage is still limited. Research conducted on same-sex couples who marry in the United States and elsewhere, has indicated that couples articulate motivations to marry that include obtaining many of the legal, social, and financial benefits of marriage (Badgett, 2009; Porche and Purvin, 2008; Lannutti, 2007). But some long-term same-sex couples have stated that marriage is unnecessary for them, given that they have established alternative methods to generate both legal and social support mechanisms (Badgett, 2009; Porche and Purvin, 2008; Kurdek, 2004; Hull, 2003). These couples articulate a feeling of being recognized by their friends and community as a couple and, accordingly, having the benefits of ‘‘couplehood’’ that might otherwise be considered the purview of married couples. Such findings mirror other research examining long-term cohabiting heterosexual couples, which has found few differences between cohabiting and married couples for those residing in countries where cohabitation is viewed as a legitimate alternative to marriage (Kiernan, 2000) or those who have plans to marry (Brown, 2000; Brown and Booth, 1996). These findings regarding both same-sex and heterosexual cohabiting couples suggest that perceived stability within a cohabiting relationship could produce many of the benefits of marriage. However, studies comparing stability and relationship quality indicators for those same-sex couples who are cohabiting to those in marriages or civil unions indicate that individuals in legalized relationships fare better on these measures (Balsam et al., 2008; Rothblum, 2005; Solomon et al., 2005). Overt exclusion from legal marriage could produce differential results from choosing cohabitation over marriage. Thus, whether same-sex elders living in cohabiting households fare worse than different-sex married elders could depend in part on the manner in which same-sex couples view their cohabiting relationship and the institutional discrimination that limits their available choices. In addition to differing views on cohabitation and marriage, outcomes for same-sex elders could vary from those of different-sex partners due to unique gender dynamics within the households. Research indicates that some marital benefits have been associated with the gendered division of labor within and outside of the household, with men typically experiencing a greater advantage from the union than women (Waite, 2005; Ross et al., 1990; Becker, 1985). For example, married men earn more than cohabiting men, due perhaps in part to their ability to ‘‘specialize’’ in the making of money and the support they receive from their wives in the household (Grossbard-Shechtman, 1993; Becker, 1985). Married women, on the other hand, are more likely to enter and exit the workforce for childcare than are nonmarried women, having a negative effect on their lifetime earnings (Becker, 1985). In contrast, research suggests that same-sex couples are less likely to divide work and household tasks according to gender norms. Rather, partners often negotiate workloads in a more egalitarian fashion, with each taking on both traditionally masculine and traditionally feminine obligations (Solomon et al., 2004; Peplau and Beals, 2004; Kurdek, 2004). It is unclear what the effects of such arrangements might be on the lifetime earnings and related benefits of same-sex elders. Further, research on marriage indicates that married men and women experience different health and longevity benefits from marriage. Data suggest that married persons, men in particular, adjust risk-taking behaviors more so than nonmarried persons (Waite and Gallagher, 2000; Ross et al., 1990). Findings indicate that altered health and longevity outcomes are due to more than just selection effects into marriage, as even sick persons who are married live longer than do those who are cohabiting, and persons who marry experience health improvements upon marriage (Umberson, 1992, 1987). It is unclear, however, whether same-sex couples would receive the same health benefit from marriage. Women are more likely to visit doctors for illnesses than are men, and some research suggests that married men’s health advantage is due in part to encouragement from their wives to schedule doctor visits (Umberson, 1987). Accordingly, the gendered benefits of heterosexual marriage could play out differently for same-sex couples, altering the outcomes for one or both of the partners. 2.3. Economic and health outcomes for same-sex cohabitors Research does indicate some differences across economic and health indicators for those in same-sex partnerships compared to different-sex partnerships. Data examining the economic status of same-sex partners compared to heterosexual cohabiting and married partners have generally suggested that same-sex cohabiting partners fall in between different-sex cohabiting and married couples in terms of wages and other economic indicators (Baumle and Poston, 2011; Baumle et al., 2009; Gates, 2009). For example, same-sex partners are less likely to own their own home than different-sex married partners, but are more likely to do so than different-sex cohabiting partners (Baumle et al., 2009). These findings are perhaps attributable in part to the fact that census data do not distinguish between same-sex cohabiting and marital partners – so some of the couples in the dataset are more like different-sex married couples, and some are more like different-sex cohabiting couples. In terms of health outcomes, there is little research that has directly examined the role that legal marriage has on the physical or mental health of same-sex partners. However, Herek (2006) argues that, given the known health benefits experienced from heterosexual marriage, it is reasonable to conclude that legally recognized marriages would result in positive health effects for gay men and lesbians. In particular, he notes that both the stability and social support derived from marriage likely result in positive mental health outcomes. Along these lines, one study using an Internet-based sample found that same-sex partners who had legally recognized partnerships or marriages demonstrated lower levels of depression and stress than those in long-term cohabiting relationships without legal recognition (Riggle et al., 2010). An important population-based study found similar results. Wight and colleagues (2012) found that lesbian, gay, and bisexual married persons experienced less

A.K. Baumle / Social Science Research 43 (2014) 60–73

63

psychological distress than did their counterparts who were not in a legally recognized relationship. Further, legally recognized relationships appeared to reduce health disparities between lesbian, gay, and bisexual persons and heterosexuals. More generally, research indicates that the availability of social support via a spouse offers physical and mental health benefits (IOM 2011; Masini and Barrett 2008; Grossman 2006). Data indicate that gay and lesbian elders are less likely to have a current cohabiting or marital partner than are heterosexual elders (MetLife, 2010; Adelman et al. 2006; Grossman et al. 2000); thus, they might be more likely to experience financial and health repercussions due to the lack of a partner in general, and a legally-recognized spouse in particular. In addition to health implications derived from social support and pooled financial resources, relationship status also has consequences for access to employer health insurance and other spousal health benefits (Cahill et al., 2000; Ash and Badgett, 2006). This lack of coverage generates health effects for gay and lesbian individuals throughout the life course. 3. Methods Drawing on the 2008–2010 pooled American Community Surveys, I analyze demographic, health, and economic characteristics of elderly same-sex cohabiting partners. In 1990, the ‘‘unmarried partner’’ response was added to other options (husband, wife, natural-born son/daughter, roommate, etc.) on the U.S. Census question asking about the relationship between the householder and other household members (Simmons and O’Connell, 2003). The ‘‘unmarried partner’’ response allows for the identification of persons in the household who have a ‘‘marriage-like’’ relationship with the householder. Census procedures allow respondents to check the ‘‘unmarried partner’’ category regardless of sex or the sex of the householder, resulting in data on ‘‘same-sex unmarried partners’’ (Baumle et al., 2009; Black et al., 2000; Simmons and O’Connell, 2003; Walther and Poston, 2004; Gates and Ost, 2004).2 Although the census data are generally accepted as useful for examining same-sex partners as a whole, some have cautioned that the data could be more problematic when used to examine subsets of the gay and lesbian population (see e.g., Gates and Ost, 2004; Baumle, 2010). Within these subsets, measurement error could be magnified, particularly if the subpopulation is one that could be prone to error. Critics of the U.S. Census numbers have long noted that the same-sex cohabiting partner data undoubtedly include some heterosexual couples, given that some individuals incorrectly mark their sex on the census forms (see e.g. Gates and Steinberger, 2009). While sex has a low rate of miscoding, small errors made within the larger population can dramatically increase the number of same-sex cohabiting partners captured in the census data (O’Connell and Feliz, 2011; Gates and Steinberger, 2009). When examining a subset of the same-sex cohabiting partner population, such as the elderly population, these concerns regarding measurement error are heightened. Measurement error concerns affect how demographers typically define their samples of same-sex partners, as described in the next section. 3.1. Sample For these analyses, I first divided couples across three primary types: same-sex cohabiting partners, different-sex cohabiting partners, and different-sex married partners. Same-sex partners are compared to both different-sex cohabiting individuals and different-sex married individuals. Given that the Census Bureau allocates same-sex couples who identify as married into the unmarried partner category,3 the released data contain some same-sex cohabitors who are likely more similar to the married different-sex partners, and others who are closer to cohabiting different-sex partners (Baumle et al., 2009). It is important to bear in mind that this is a cross-sectional analysis of the characteristics of currently partnered individuals. Accordingly, the data cannot speak extensively to the existence or duration of prior relationships or the duration of the current relationship. Out of these same-sex and different-sex couples, I followed the suggested practice by Gates and Steinberger (2009) and others of dropping all individuals in the dataset whose marital status was allocated by the Census Bureau and whose sex is nonverified via ACASI procedures. There is an ongoing debate among demographers regarding whether the inclusion of individuals with allocated marital status is appropriate for analyses of same-sex partners (see e.g. Black et al., 2000; O’Connell and Gooding 2007; Baumle et al., 2009; Gates and Steinberger, 2009; O’Connell and Feliz, 2011). By including these individuals, some heterosexual married couples with a miscoded sex will be incorporated into the same-sex unmarried partner category. By excluding these individuals, however, those who are in same-sex partnerships and intentionally identified themselves as married on the census are not included in the same-sex partner analysis. This perhaps biases results so that same-sex cohabiting partners look less like married heterosexual couples. Nonetheless, for the purposes of these analyses, I have followed work by other demographers (Gates and Steinberger, 2009) by taking the more conservative approach and, in cases where sex is not verified through ACASI, including only those individuals whose marital status was not allocated. 2 Although the census data provide a large dataset for the analysis of issues related to cohabiting same-sex partners, the data are nonetheless limited in several respects. Most notably, the census data only capture gay men and lesbians who choose to identify as unmarried partners, which does not include those who are single or who do not self-identify. Further, researchers have explored how well the census data have enumerated the actual numbers of same-sex partners in the United States, the validity of the geographic variation in prevalence rates, and the extent to which there is error based on miscoding of the sex variable (Baumle et al., 2009; Badgett and Rogers, 2003; Black et al., 2000). Through comparisons with other representative datasets, along with employing various methods of assessing validity and sampling error, these concerns have been explored in depth by those using the census data to study same-sex partners and will not be addressed in this article (see e.g., Baumle et al., 2009; O’Connell and Gooding, 2007; Baumle et al., 2009; Fields and Clark, 1999). 3 This allocation is made in response to requirements of the Defense Against Marriage Act (Simmons and O’Connell (2003).

64

A.K. Baumle / Social Science Research 43 (2014) 60–73

In the analyses reported below, I examine demographic, health, and economic characteristics of elders in same-sex cohabiting relationships, making comparisons to different-sex couples, married and cohabiting. All analyses of household and/or couple characteristics are calculated using a sample that includes both members of couples when both individuals are 65 or older. Analyses of the characteristics of individual elders who are members of same-sex or different-sex couples are restricted to those partners who are 65 years of age and older. Whether couples or individuals are used in the analyses is indicated in the data tables.

3.2. Variables and analyses Due to the dearth in the current literature on demographics of elder same-sex cohabiting partners, I have focused partly on bivariate analyses within this article, noting whether differences between older individuals in same-sex cohabiting couples and different-sex couples are statistically significant. As detailed below, I also run logistic regression analyses to analyze differences in health outcomes across couple types. I apply the appropriate sample adjustment methods, using Stata’s ‘‘svy’’ methods, to adjust my analyses according to population weights.4 For individuals 65 and over, I first provide context by presenting characteristics that address general demographics. In particular, I examine: age distributions, with age measured in years; sex (male and female); and race (coded across six categories: White, African American, American Indian, Asian/Pacific Islander, Other, and Two or more Races). In logistic regression analyses, I instead use a dummy variable for race, coded as 0 for white and 1 for nonwhite. In addition to these basic demographics, I also examine several variables related to household composition. I analyze the number of persons in the household (the number of person records reported for each household) and the number of ‘‘own’’ children reported for each individual (which includes biological, step, or adopted children). I further explore the children question by including two dummy variables. One examines whether a young (likely dependent) child is in the household; a dependent child is defined as a child under the age of 24, per the IRS regulations that allow individuals to claim children under this age who are in college as dependents (coded as 0 for none, and 1 for one or more children under the age of 24). I also include a dummy variable assessing the presence of older children in the household, i.e. 24 and over (0 indicates none, and 1 indicates one or more children 24 and over). And, finally, I examine whether individuals in any of the household types reported prior marriages, focusing on individuals who reported that they were ‘‘never married’’ (coded as 0) as compared to those who reported a current or prior marriage (married, divorced, or separated, coded as 1). I then explored two substantive outcomes for same-sex cohabiting partners compared to different-sex cohabiting and married partners: economic and health outcomes. At the household level, I explored several economic characteristics that are particularly relevant for elders in terms of current income streams from wages, retirement sources, and public assistance.5 I examine household income from all sources, income from wages (total pre-tax wage and salary income earned from employers in the previous year), income from Social Security (pre-tax income received from Social Security benefits, pensions, survivor benefits, or disability benefits), income from retirement sources (pre-tax retirement, survivor, and disability income other than from Social Security), income from Supplemental Security Income (pre-tax income from SSI), and income from welfare (pre-tax income from public assistance programs). In addition, I included a variable to assess whether both members of a couple were in the labor force (coded 0 for no, and 1 for yes), and a variable indicating whether the home is owned (coded 0 for no, and 1 for yes). In terms of health characteristics, I report both bivariate analyses and results from logistic regression analyses. These analyses occur at the individual level, rather than couple level, meaning that all analyses focus only on partners who are 65 or older. In assessing health outcomes, I am necessarily limited to those variables captured by the ACS. Specifically, I include dummy variables for several reported disabilities (coded 0 for not reported, and 1 for reported): cognitive, ambulatory, independent living, self care, vision, and hearing difficulties. Further, I was able to assess differences in health insurance across couple type through several dummy variables (coded 0 for not reported, and 1 for reported): possess any health insurance coverage, Medicaid coverage, Medicare coverage, and VA coverage. The logistic regression analyses further explore the manner in which health outcomes might be influenced by relevant individual-level variables. These analyses are conducted in order to explore the characteristics that differ significantly between older individuals in same-sex cohabiting or different-sex cohabiting or marital couples; they are not conducted in order to predict an older person’s odds of being in a same-sex or different-sex relationship (see e.g. Poston and Baumle, 2010; Bogaert, 2004). For each sex, I estimate two logistic regression equations at the individual-level: (1) one compares differentsex cohabiting partners and same-sex cohabiting partners, and (2) another compares different-sex married partners and same-sex cohabiting partners. The dependent variable is a dichotomous dummy variable for relationship type, with 0 indicating a different-sex relationship and 1 indicating a same-sex relationship. I then include fifteen independent variables, measured at the individual-level, including: age, being nonwhite, presence of a young child, presence of an older child, 4 Given the disparities in sample size across my groups (see Ns reported in tables), I took two measures to correct for possible biases. Given that unequal sample sizes can increase the likelihood of unequal variances, thus violating assumptions for student’s t-test, I have used t-tests for samples with unequal variances. In addition, I have taken subsamples (10% of cohabiting heterosexuals and 1% of married heterosexuals) and run the bivariate comparisons and logistic regression analyses with the subsamples. Although, as would be expected, some of the means change slightly when using the subsamples, the inferences remain the same regarding statistical significance. 5 When using the pooled 2008–2010 ACS data, all dollars are standardized to 2010 dollars.

65

A.K. Baumle / Social Science Research 43 (2014) 60–73 Table 1 Age characteristics for different-sex and same-sex partners 65 and over (individuals).

*

Different-sex married partners

Different-sex cohabiting partners

N (Individuals)

Male: 430,337 Female: 334,867

Male: 9845 Female: 6855

Same-sex cohabiting partners Male: 986 Female: 780

Mean age

73

71.9

71*+

Age categories 65–69 70–79 80–89 90–99

Percentage in age category 37.4% 45.3% 16.1% 1.2%

Percentage in age category 45.5% 41.1% 12.1% 1.3%

Percentage in age category 51.6% 39.0% 7.8% 1.6%

Mean age: Men

73.4

72.2

70.9*+

Age categories: Men 65–69 70–79 80–89 90–99

Percentage in age category 35.5% 45.3% 17.7% 1.6%

Percentage in age category 44.6% 40.9% 13.1% 1.4%

Percentage in age category 51.4% 40.6% 7.2% .8%

Mean age: Women

72.5

71.6

71.2*

Age categories: women 65–69 70–79 80–89 90–99

Percentage in age category 39.9% 45.4% 14.0% .7%

Percentage in age category 46.8% 41.5% 10.7% 1.1%

Percentage in age category 51.9% 37.1% 8.5% 2.5%

Statistically significant from different-sex married partners, at or below the .05 level of significance. Statistically significant from different-sex cohabiting partners, at or below the .05 level of significance.

+

having been previously married, number of persons in the household, cognitive difficulty, ambulatory difficulty, independent living difficulty, self-care difficulty, vision difficulty, hearing difficulty, Medicaid insurance coverage, Medicare insurance coverage, and VA insurance coverage. The variables are measured as described above, and were selected based on results from the bivariate analyses as well as diagnostic tests for collinearity. I exponentiate the regression logit coefficients and report the odds ratios in my results.

4. Results 4.1. General demographics Given the lack of data on elders in same-sex cohabiting partnerships, I first provide an overview of general demographic characteristics of individuals in this couple type and how they compare with those in different-sex cohabiting and married households. The data indicate that individuals 65 and older who identified as same-sex cohabitors are, on average, one year younger (at 71) than different-sex cohabiting partners (71.9) and two years younger than different-sex married partners (73); these differences are statistically significant at the .05 level (Table 1). When examined by age categories, there is a greater proportion of same-sex cohabitors than different-sex partners in the lower age category of 65–69 (Table 1), whereas there is a greater proportion of different-sex partners than same-sex cohabitors in the upper age category of 80–89. Looking at the mean age by sex, it appears that the age difference across couple types is more pronounced for men than for women. For men, for example, there is over a two year difference between the mean age of men in different-sex married partnerships and men in same-sex cohabiting partnerships (Table 1). For women, there is only a 1.3 year difference. In Table 2, I examine sex and race differences by relationship type. The data reflect that across all couple types, a greater proportion of individuals 65 and older are male than are female. Table 2 also reveals that individuals over the age of 65 who are in same-sex cohabiting partnerships are less racially diverse than those in different-sex partnerships. Approximately 93% of elderly individuals in same-sex cohabiting partnerships are white, compared to 90% of elderly individuals in different-sex married partnerships and 86% of those in different-sex cohabiting partnerships. The most notable difference across racial groups is in the African American category, with 3.6% of elderly same-sex cohabitors being African American, as compared to 5.3% of different-sex married individuals and 8.9% of different-sex cohabitors. I also examined differences in household composition across the three relationship types (Table 3). I first assess household size by looking at the total number of persons living in the household with a partner age 65 and older. Results show that all relationship types have between 2.1 and 2.4 individuals in the household, with male same-sex cohabitors reporting the smallest household size and female same-sex cohabitors reporting the largest (Table 3). Looking more closely at those individuals within the household, Table 3 reflects comparisons in terms of the presence of children. Approximately 10% of

66

A.K. Baumle / Social Science Research 43 (2014) 60–73

Table 2 Sex and race characteristics for different-sex and same-sex partners 65 and over (individuals).

*

Different-sex married partners

Different-sex cohabiting partners

Same-sex cohabiting partners

N (Individuals) Sex Male Female

Male: 430,337 Female: 334,867 Percentage in sex category 56.7% 43.3%

Male: 9845 Female: 6855 Percentage in sex category 59.8% 40.2%

Male: 986 Female: 780 Percentage in sex category 55.5%+ 44.5%+

Race White African American American Indian Asian/Pacific Islander Other Two or more races

Percentage in race category 89.6% 5.3% .4% 3.0% 1.0% .7%

Percentage in race category 86.2% 8.9% .7% 1.3% 1.8% 1.1%

Percentage in race category 92.8%*+ 3.6% .6% 1.5% .4% 1.1%

Statistically significant from different-sex married partners, at or below the .05 level of significance. Statistically significant from different-sex cohabiting partners, at or below the .05 level of significance.

+

Table 3 Household composition for different-sex and same-sex partners 65 and over (individuals). Different-sex married partners

*

Different-sex cohabiting partners

Same-sex cohabiting partners

Male: 9845 Female: 6855

Male: 986 Female: 780

N (Individuals) Male: 430,337 Female: 334,867 Mean number of persons in household Overall 2.3 Male 2.3 Female 2.2

2.2 2.3 2.2

2.2 2.1*+ 2.4*+

Mean number of family members in household Overall 2.2 Male 2.3 Female 2.2

1.1 1.1 1.1

1.2* 1.1* 1.3*+

Percentage with own child under 24 in household Overall 1.2% Male 1.9% Female .3%

1.1 1.7% .14%

.3%* .4%* .2%

Percentage with own child 24 or older in household Overall 10.4% Male 10.7% Female 9.9%

3.4% 2.6% 4.7%

4.7%*+ 1.8%* 8.4%+

Percentage never married Overall Male Female

6.7% 7.6% 5.2%

55.5%*+ 66.5%*+ 41.7%*+

0% 0% 0%

Statistically significant from different-sex married partners, at or below the .05 level of significance. Statistically significant from different-sex cohabiting partners, at or below the .05 level of significance.

+

different-sex married partners who are 65 and older reported having a child age 24 or older in the household (perhaps as a dependent, or perhaps as a caretaker). In comparison, approximately 8% of females 65 and older in a same-sex cohabiting relationship, 3.4% of different-sex cohabitors, and only 1.8% of men in a same-sex cohabiting relationship have an older child in the household. As would be expected, the presence of younger ‘‘own children’’ in the household is limited for all couple types, although a smaller proportion of female and male same-sex cohabitors have young children in the household. These results suggest differences in terms of both dependents and potential caretakers across elder households. Finally, I examine whether individuals report having been previously married; in particular, whether individuals indicate that they have ‘‘never married.’’ As reflected in Table 3, about two-thirds of men 65+ in a same-sex cohabiting relationship report having never married, and about 42% of all females in a same-sex cohabiting relationship. This is in contrast to less than 7% of those in different-sex cohabiting partnerships.

4.2. Economic outcomes Turning to the first of the outcome variables, I examine several indicators of economic well-being across the couple types. Specifically, the ACS data permit examination of current income sources, including income from wages, Social Security, Supplemental Support Income, and other public assistance. In addition, I assess retirement resources, including reported income

67

A.K. Baumle / Social Science Research 43 (2014) 60–73 Table 4 Financial support for different-sex and same-sex couples with both partners 65 and over (household).

N (Households) Mean household income, all sources Mean household income from wages Percentage with dual employment Percentage receive income from SSI Mean household income from SSI Percentage receive income from welfare Mean household income from welfare Percentage receive income from social security Mean household income from social security Percentage receive income from retirement Mean household retirement income Percentage owning home

*

Different-sex married partner households

Different-sex cohabiting partner households

Male same-sex cohabiting partner households

Female same-sex cohabiting partner households

315,318 $68,440

5173 $65,859

224 $94,601*+

168 $77,647*+

$40,610

$35,449

$49,089

$33,048

5.4%

6%

6.7%

9.6%*+

1.8%

3.2%

2.0%

1.5%

$9935

$7497

$9812

$11,394+

.4%

.7%

1.5%*

.8%

$5811

$4713

$7029

$5957

90.9%

90.4%

90.2%

92.6%

$22,488

$24,030

$24,833*

$25,107*

38%

39%

39%

57%*+

$25,765

$22,427

$34,948*+

$31,293*+

91%

85%

91%+

92%+

Statistically significant from different-sex married partners, at or below the .05 level of significance. Statistically significant from different-sex cohabiting partners, at or below the .05 level of significance.

+

from retirement sources and home ownership. The data indicate a mixed picture for older same-sex cohabiting couples in terms of financial resources. Table 4 depicts the results for measures of economic well-being. All of the variables are reported for couples where both individuals are 65 and over.6 I first examine differences in total household income. This includes all pre-tax income received by the partners in the household. Male same-sex cohabiting couples had the highest reported household income at $94,601, followed by female same-sex cohabiting couples at $77,647, different-sex married couples at $68,440, and different-sex cohabiting couples at $65,859; the household incomes for male and female same-sex couples are significantly different from those of both different-sex married and different-sex cohabiting households. Data indicating the mean income from wages reflect a more typical pattern seen in literature on sexual orientation and income: male same-sex cohabiting couples earn the most from wages, followed by different-sex married couples, different-sex cohabiting couples, and then female same-sex cohabiting couples. These differences, while large, are not statistically significant at the .05 level, although the wages of male same-sex cohabiting households are significantly different from those of different-sex couples at the .10 level. In addition to mean income from wages, I considered whether both partners in the household were still in the labor force at age 65 or older. There were no statistically significant differences between dual labor force participation for male same-sex cohabiting couples as compared to different-sex married or different-sex cohabiting couples; approximately 5–7% of all these couple types reported dual labor force participation. Approximately 10% of female same-sex couples, however, reported dual labor force participation; this was significantly different from the participation of different-sex married and cohabiting partners. Although some income is derived from wages, elderly couples are also dependent on retirement funds (social security and personal savings) and public assistance programs for financial support. As depicted in Table 4, ACS data indicate that 90–92% of individuals in all couples types received income from Social Security, with females in same-sex cohabiting couples reporting a slightly higher percentage than the other couple types. Further, different-sex married couples report a slightly lower mean income from Social Security than do the other couple types (Table 4). Looking to other retirement income (pre-tax retirement, survivor, and disability income other than from Social Security), male same-sex cohabiting couples and both types of different-sex couples reported about 39% receiving retirement income 6 I chose to report couple income, rather than solely that of individuals, given the strong relationship between financial resources and marital status. For example, some Social Security benefits are available only to spouses; qualification for other public benefits is often calculated based on income within a marriage; etc. Further, I focused on couples where both individuals are 65 and over given that some benefits are generally unavailable to individuals until they reach the age of 65 (e.g. Social Security, Medicare), and more individuals will have retired by the age of 65 (making comparisons across couple types more comparable). Finally, all mean incomes were calculated for those reporting that particular type of income; in other words, values of 0 were excluded from the calculation. For some types of income, values of 0 varied notably across couple type; thus, including these 0s skewed the means and reduced comparability. The reported proportions receiving each income type, however, capture the variability in receipts across couple types.

68

A.K. Baumle / Social Science Research 43 (2014) 60–73

in the prior year. Approximately 57% of female same-sex cohabiting couples, however, reported retirement income; this is significantly different from different-sex married and cohabiting couples at the .05 level. Both male and female same-sex cohabiting couples report significantly higher mean retirement income than do different-sex married or cohabiting couples. For example, the mean household retirement income reported by male same-sex cohabiting couples exceeds that of different-sex married couples by almost $9200, and the mean household retirement income for female same-sex cohabiting couples exceeds that of different-sex married couples by about $5500. Approximately 2–3% of all couple types receive Supplemental Support Income (‘‘SSI’’), with no statistically significant differences. Of those receiving SSI, however, the mean income received by female same-sex cohabiting couples is significantly larger than that received by different-sex cohabiting couples (although not than that received by different-sex married couples). Looking to income from welfare and other state and local public assistance programs, the percentage of male same-sex cohabiting couples receiving public assistance (1.5%) is about three times greater than that of different-sex married couples (.4%); this difference is statistically significant at the .05 level. Home ownership serves as another indicator of financial resources and retirement support, and is often one of the primary sources of saving for couples. Compared to different-sex married couples, there were no significant differences in reported home ownership for same-sex cohabiting couples, with approximately 91% reporting owning their own home. Different-sex cohabiting partners were, however, significantly less likely to own their home than any of the other couple types. 4.3. Health outcomes The second outcome variable involves several measures of health. The census permits the examination of six measures of disability: cognitive, ambulatory, independent living, self-care, vision, and hearing difficulties. Table 5 presents the percentage of elderly individuals in each of the couple types who reported one of these disabilities. Given the relationship between sex and many of these health indicators, I report separate results for males and females within each couple type. Further, given the influence of individual characteristics on health, both bivariate results and logistic regression analyses are presented to analyze the health variables. 4.3.1. Bivariate results The general pattern indicates that elderly individuals in male same-sex cohabiting partnerships have lower percentages of reported disabilities than do the other couple types, whereas elderly individuals in female same-sex cohabiting partnerships trend toward higher percentages of reported disabilities. Ambulatory difficulty was reported the most by all couple types at about 17% of all individuals, and was significantly higher for females in same-sex cohabiting partnerships (22% reporting) and significantly lower for males in same-sex cohabiting partnerships (13% reporting). Hearing difficulty was the next most prominent disability, and is reported particularly by men in different-sex relationships but significantly less so by men in same-sex cohabiting relationships. On the flip side, a significantly greater percentage of females in same-sex cohabiting partnerships reported hearing difficulty at 10%, than females in different-sex relationships, at about 7 to 8%. Finally, a statistically significant, lower percentage of males in same-sex cohabiting relationships (4.1%) reported cognitive difficulty than males in different-sex relationships (6–6.5%). In addition to measures of disabilities, the ACS also permits the examination of whether an individual has healthcare coverage from any source, as well as the specific sources of coverage. As indicated in Table 5, findings suggest that members of all couple types have access to health care coverage, but the type of coverage varies to some degree by couple type. Both males and females in same-sex cohabiting couples are less likely than different-sex cohabiting or married couples to report Medicare coverage. Females in same-sex cohabiting couples are more likely to report Medicaid coverage, indicating income or health disparities. And males in same-sex cohabiting couples are less likely to report VA coverage, while females are more likely to report VA coverage, perhaps reflecting differences in military service by sexual orientation. 4.3.2. Logistic regression results In order to examine whether these health differences hold after controlling for individual characteristics, I conducted logistic regression analyses. As indicated in Section 3, I conduct these analyses only to describe characteristics of older samesex cohabiting partners versus different-sex cohabiting and married partners, rather than to predict the odds of a person being in a same-sex versus different-sex relationship. The second column in Table 6 depicts the results for older men in same-sex cohabiting relationships compared to those in different-sex married relationships. For general demographic characteristics, results indicate that same-sex cohabiting partners are younger and whiter than different-sex married partners. In terms of health indicators, the difference in hearing remains statistically significant, with those who are in a same-sex cohabiting relationship being less likely to have difficulty hearing than those in a different-sex marriage. In the comparison of older male same-sex cohabiting partners with older males in different-sex cohabiting partnerships, we see very similar patterns in terms of the characteristics that are significantly different: same-sex cohabiting partners are younger, whiter, and have less hearing difficulty than different-sex cohabiting partners. Further, they are less likely to have cognitive difficulty than are different-sex cohabiting partners; the odds of those with cognitive difficulty being in a same-sex relationship versus a different-sex relationship are 43% less than those who do not have cognitive difficulty. Those who have difficulty with independent living, however, are about two times more

69

A.K. Baumle / Social Science Research 43 (2014) 60–73 Table 5 Health indicators for different-sex and same-sex partners 65 and over (individuals).

N (Individuals) Cognitive difficulty Overall Male Female

Different-sex cohabiting partners (%) Male: 9845 Female: 6855

Same-sex cohabiting partners (%) Male: 986 Female: 780

5.9 6.5 5.1

5.8 6.0 5.5

4.6* 4.1*+ 5.4

16.9 16.5 17.4

17.7 17.2 18.5

16.7 12.6*+ 21.9*+

Independent living difficulty Overall 9.7 Male 9.3 Female 10.3

8.8 7.7 10.4

10.1 8.9 11.6

Ambulatory difficulty Overall Male Female

*

Different-sex married partners (%) Male: 430,337 Female: 334,867

Self-care difficulty Overall Male Female

5.4 5.4 5.4

5.1 4.8 5.5

5.2 3.9 6.8

Vision difficulty Overall Male Female

4.6 4.9 4.2

5.3 5.5 5.0

4.5 4.4 4.7

Hearing difficulty Overall Male Female

13.6 18.4 7.3

13.4 17.2 7.7

9.2*+ 8.4*+ 10.2*+

Any health care Overall Male Female

99.6 99.6 99.6

98.9 98.8 99.1

99.4 99.7+ 98.9*

Health insurance through medicaid Overall 9.6 Male 9.6 Female 9.7

14.9 14.0 16.2

12.0*+ 10.8+ 13.4*

Health insurance through medicare Overall 97.2 Male 97.1 Female 97.3

95.7 95.7 95.7

95.0*+ 94.6*+ 95.6*

Health insurance through VA Overall 9.9 Male 16.6 Female 1.1

12.6 20.1 1.4

9.4+ 14.0*+ 3.7*+

Statistically significant from different-sex married partners, at or below the .05 level of significance. Statistically significant from different-sex cohabiting partners, at or below the .05 level of significance.

+

likely to be in a same-sex cohabiting relationship versus a different-sex cohabiting relationship compared to those who do not have difficulty with independent living. Finally, results indicate that males in same-sex cohabiting partnerships are less likely to be on Medicaid than are different-sex cohabiting partners, and are less likely to have health insurance through the VA. Like male same-sex partners, results indicate that female same-sex cohabiting partners are younger than female different-sex married partners, and that households are somewhat larger and less likely to include an older child (Table 6). Health indicators support the earlier assessment that females in same-sex cohabiting relationships have greater health disabilities. Specifically, the odds of those individuals with ambulatory difficulty and hearing difficulty being in a same-sex cohabiting relationship versus a different-sex married relationship are significantly greater than those without those disabilities. Perhaps corresponding with these heightened disabilities, results indicate that the odds of those who have Medicaid insurance being in a same-sex cohabiting versus different-sex married relationship are 43% greater than those who do not have Medicaid insurance. And results indicate that those older women who have VA health insurance are more than three times as likely to be in a same-sex cohabiting versus a different-sex married relationship than those without VA insurance. In the comparison of older females in same-sex cohabiting relationships with those in different-sex cohabiting relationships, there are similar patterns in terms of the following: same-sex cohabiting partners are younger, have slightly larger households, have more ambulatory and hearing difficulties, and are more likely to have VA health insurance. Unlike in

70

A.K. Baumle / Social Science Research 43 (2014) 60–73

the comparison with females in married different-sex couples, however, female same-sex cohabiting couples do not have significantly higher odds of being on Medicaid than females in different-sex cohabiting relationships.

5. Discussion and conclusion Prior research on different-sex couples indicates that marriage can provide cumulative economic and health benefits throughout the lifetime that are not found in cohabitation (Waite, 2005; Seltzer, 2004; Brown, 2000; Horwitz and White, 1998; Waite and Gallagher, 2000). Due to the relatively recent access to marriage for same-sex couples, it remains unclear whether the same lifetime benefits would accrue to same-sex couples. The analyses presented herein provide preliminary support for the proposition that elder same-sex cohabiting couples differ in important ways from both different-sex cohabiting and different-sex married couples. These results provide insight into the manner in which the sex of one’s partner might produce different economic and health outcomes for married and cohabiting elders. Notably, about two-thirds of men and over 40% of women who are 65+ in same-sex partnerships report having never married. This is an important finding in that it is a validity check in terms of supporting that these data are capturing the intended population. There is a clear difference on this variable for different-sex and same-sex couples, supporting the notion that those who have identified as same-sex cohabitors are likely to be in a true same-sex relationship (rather than selecting the category in error). Second, and importantly for this analysis, this finding suggests that individuals in samesex cohabiting relationships have often lived their lives outside of heterosexual legal marriage. Accordingly, findings from these analyses can be interpreted as reflecting (for many) the results of a life lived without legal marriage. As indicated by the marital institutional theory, one would expect to find that same-sex cohabiting elders demonstrate worse outcomes due to their relative exclusion from heterosexual marriage throughout their lives. These results support the notion that one of the primary benefits that might be afforded via heterosexual marriage involves the pooling of financial resources, particularly given that those resources involve the typically higher lifetime earnings of a male partner. Male same-sex cohabiting partners appear to fare well on both economic and health outcomes as compared to their differentsex cohabiting counterparts. Indeed, they fare better on many outcomes than do different-sex married men, contrary to the hypothesized expectations. Results concerning elder females in same-sex cohabiting partnerships, however, suggest that these women might incur a cumulative disadvantage over their lifetime from the lack of a male partner’s income. A greater proportion of elderly female same-sex cohabiting households are engaged in dual labor force participation than are different-sex married or different-sex cohabiting households. This indicates that females in same-sex cohabiting relationships are more likely to be working into older ages, perhaps in order to further build their retirement resources or due to added financial obligations produced by their greater reported health problems. Further, their mean household income from wages is lower than that of differentsex married or cohabiting couples – even though they have a greater proportion reporting income derived from two wage earners. This finding suggests that female same-sex cohabitors might be playing ‘‘catch up’’ when they enter their elder years, perhaps as a consequence of not having a (typically) higher-earning male partner. In terms of retirement income, results indicate that female same-sex cohabitors are relying more on retirement income than are other couple types. Further, female same-sex cohabiting couples have a higher mean household income from retirement than any couple type except male same-sex cohabiting couples. These data do not permit an assessment of the overall amount of retirement resources accumulated by couple types; thus, a higher current receipt of retirement income does not necessarily equate to greater overall resources. It is possible that female same-sex cohabitors are relying more heavily on their retirement savings than are other couple types at present, but could exhaust their resources more quickly. The findings regarding lower wage earnings and greater dual labor force participation seem to support the notion that females in samesex cohabiting relationships have greater need to draw on their retirement resources and continue to build those resources into older age, perhaps due to a deficit from a lifetime of lower household earnings. It is also possible that female same-sex cohabitors could have greater retirement resources than different-sex couples due to a reduced likelihood for one partner to leave the workforce in order to have children (see e.g. Peplau and Fingerhut, 2004). Older same-sex cohabiting couples appear just as likely as older married couples to own their own home, and more likely than different-sex cohabiting couples to do so. These findings indicate that same-sex cohabitors, like different-sex married persons, might be pooling assets for joint purchases, signaling a degree of commitment that could be similar to that found in marriage. These results could reflect that cohabitation means something different for same-sex couples than for different-sex couples, due to the lack of access to legal marriage as an option (see Kiernan, 2004 for similar findings regarding cohabitation in Western Europe). Looking to health outcomes, results again reflect the manner in which sex affects the experience of old age for same-sex cohabitors as compared to different-sex married or cohabiting individuals. First, female same-sex cohabitors are more likely to be disabled than individuals in different-sex marriages or cohabiting couples. In addition, females in same-sex couples report the use of Medicaid insurance coverage significantly more than do different-sex married couples, which could signal both health and income disadvantages. On the other hand, results indicate that older males in same-sex cohabiting couples report fewer health difficulties than those in different-sex married or cohabiting couples, and are also less likely to report reliance on Medicaid than do those in different-sex cohabiting couples. This apparent health advantage experienced by male same-sex cohabitors could be attributable to higher lifetime earnings, which increase access to healthcare and related po-

71

A.K. Baumle / Social Science Research 43 (2014) 60–73 Table 6 Logistic regression analyses: odds ratios for individual-level characteristics for those 65 and over.

Age Nonwhite Young child Older child Married in lifetime Persons in HH Cognitive difficulty Ambulatory difficulty Independent living difficulty Self-care difficulty Vision difficulty Hearing difficulty Medicaid insurance Medicare insurance VA insurance F-test *

Males: same-sex cohabiting compared to different-sex married

Females: same-sex cohabiting compared to different-sex married

0.93* 0.62* 0.38 0.17* Not included

0.96* 0.79 0.32 0.44* Not included

Males: same-sex cohabiting compared to different-sex cohabiting

Females: same-sex cohabiting compared to different-sex cohabiting

0.97* 0.23* 0.51 1.69 0.08*

0.98* 0.39* 0.62 1.46* 0.19*

1.36* 0.74

1.43* 0.94

0.74* 0.57*

1.27* 0.81

1.35

1.41*

0.91

1.31*

1.49

0.99

2.19*

1.04

0.88

1.12

0.85

1.13

1.61 0.57

1.16 *

1.51

1.22 *

0.65

0.99 *

1.47*

1.27

1.43*

0.72*

0.80

0.69

0.66

1.33

1.49

0.92 94.39*

*

3.35 61.26*

*

0.76 47.91*

3.32* 20.36*

Statistically significant at or below the .05 level of significance.

sitive health practices; on the flip side, the apparent health disadvantage experienced by female same-sex cohabitors could reflect lower lifetime earnings. These findings suggest that elder same-sex cohabitors might have experiences in old age that are substantively different from those in different-sex couples, but that these differential experiences speak primarily to the sex of their partner rather than their access to legal marriage. Consequently, the marital institution theory lacks strong support from the data, raising questions regarding the manner in which same-sex couples might have utilized alternative means to garner some of the protections typically associated with marriage. Perhaps some of these couples, particularly those who have lived much of their lives outside of legal marriage, have been able to draw upon financial, legal, and social support mechanisms to produce stability and a pooling of resources that mitigate against the exclusion from marriage (see similar discussion in Fredriksen-Goldsen and Muraco, 2010; Badgett, 2009; Porche and Purvin, 2008). At the same time, however, this does not indicate that legal marriage is irrelevant to the welfare of elder same-sex cohabitors. Given that heterosexual marriage carries with it access to a variety of health and financial benefits through one’s employer (e.g. health insurance) and through the federal and state governments (e.g. social security benefits), same-sex cohabitors who have not spent most of their lives in marriage have undoubtedly been affected. For female same-sex couples, lack of marriage might prove more detrimental over the course of their lives, given that they likely begin from a place of financial disadvantage compared to different-sex couples or male same-sex couples. Accordingly, access to health insurance, tax benefits, social security benefits, and other marriage-related benefits could aid in defraying the disparity in health and economic outcomes that female same-sex cohabitors exhibit. Given the U.S. Supreme Court’s recent decision striking down key provisions of the Defense of Marriage Act, changes that provide access to federal benefits for individuals who are able to obtain legal marriage could be a step toward reducing some of these financial disparities in old age (United States v. Windsor, 2013). Nonetheless, same-sex couples will continue to face challenges regarding access to health insurance and state-provided benefits which will likely affect their financial and health positions at retirement. Further, the exclusion from legal marriage carries with it harmful effects incurred as a product of institutional discrimination. Consequently, measures of welfare on other health indicators could reflect harms that the ACS data do not capture. Finally, it is important to keep in mind the limitations of this particular dataset when evaluating outcomes for elder samesex partners. Through the ACS data, comparisons can be made between those who identify on the form as married or unmarried partners. It is likely that most different-sex partners feel few, if any, identity conflicts when completing the form. For same-sex partners, on the other hand, individuals might identify as partners or as roommates or other relations (see Gates, 2010 for a discussion of identification on the census). Accordingly, the ACS only captures those partners who choose to

72

A.K. Baumle / Social Science Research 43 (2014) 60–73

identify their relationships. Some of the data raise questions as to whether sex and race play a role in identification for elders in same-sex partnerships. Results suggest that a greater proportion of older same-sex partners are male than female; this is unexpected given women’s longevity advantage. The results could signal that older gay men are more likely to identify as same-sex cohabitors on the census than are older lesbians, thus generating the appearance of a greater proportion of male same-sex partners. Further, the data suggest that a greater proportion of elders in same-sex partnerships are white than are different-sex partners. This finding is consistent with the notion that identification as gay or lesbian (as implied by same-sex partnerships) has historically been more connected with the white gay and lesbian community (Fredriksen-Goldsen and Muraco, 2010; Herdt and de Vries, 2004). For older same-sex cohabitors, then, one might expect to see less diversity than is reflected in the general population due to this cohort effect. It is possible that those individuals who live a more closeted life are less likely to identify their relationship, and perhaps more likely to experience financial and health repercussions. Future research on identification of elders in same-sex partnerships could perhaps shed light on whether and how the ACS data are affected by the particular pool of elders electing to identify. In addition, it must be reiterated that these data provide a cross-sectional picture of the population of elderly partnered individuals in different-sex or same-sex relationships. Details are lacking regarding how long individuals have been partnered, whether this is their first partnership, or (for those who have been married) the length of time they have spent in heterosexual marriages. Accordingly, additional research on lifetime relationship patterns is needed in order to evaluate the cumulative effects of same-sex partnerships for elders. References Adelman, M., Gurevich, L., de Vries, B., Blando, J., 2006. Openhouse: Community building and research in the LGBT aging population. In: Kimmel, D., Rose, T., David, S. (Eds.), Lesbian, Gay, Bisexual, and Transgender Aging: Research and Clinical Perspectives. Columbia University Press, New York. Ash, M., Badgett, M.V.L., 2006. Separate and Unequal: the effect of unequal access to employment-based health insurance on same-sex and different-sex couples. Contemporary Economic Policy 24 (4), 582–599. Bachrach, C., 1987. Cohabitation and reproductive behavior in the U.S. Demography 24, 623–637. Badgett, M.V.Lee, 2009. When Gay People Get Married: What Happens When Societies Legalize Same-Sex Marriage. NYU Press, New York, NY. Badgett, M.V. Lee, Rogers, Marc A., 2003. Left Out of the Count: Missing Same-sex Couples in Census 2000. Amherst, MA: Institute for Gay and Lesbian Strategic Studies. Balsam, K.F., Beauchaine, T.P., Rothblum, E.D., Solomon, S.E., 2008. Three-year follow-up of same-sex couples who had civil unions in vermont, same-sex couples not in civil unions, and heterosexual married couples. Developmental Psychology 44, 102–116. Baumle, Amanda K., 2010. Border Identities: intersections of ethnicity and sexual orientation in the U.S. – Mexico Borderland. Social Science Research 39, 231–245. Baumle, Amanda K., Poston, Dudley L., 2011. The economic cost of being homosexual: a multilevel analysis. Social Forces 89, 1005–1031. Baumle, Amanda K., Compton, D’Lane, Poston Jr., Dudley L., 2009. Same-Sex Partners: The Social Demography of Sexual Orientation. SUNY Press, New York. Becker, Gary S., 1985. Human capital, effort, and the sexual division of labor. Journal of Labor Economics 3 (1), 33–58. Black, D., Gates, G.J., Sanders, S.G., Taylor, L., 2000. Demographics of the gay and lesbian population in the United States: Evidence from available systematic data sources. Demography 37 (2), 139–154. Bogaert, A.F., 2004. Asexuality: Prevalence and Associated Factors in a National Probability Sample. The Journal of Sex Research 41 (3), 279–287. Brines, J., Joyner, K., 1999. The ties that bind: principles of cohesion in cohabitation and marriage. American Sociological Review 64, 333–355. Brown, S.L., 2000. The effect of union type on physiological well-being: depression among cohabitors versus marrieds. Journal of Health and Social Behavior 41, 241–255. Brown, S., Booth, A., 1996. Cohabitation versus marriage: a comparison of relationship quality. Journal of Marriage and the Family 58, 668–678. Bumpass, L.L., Lu, H., 2000. Trends in cohabitation and implications for children’s family contexts in the United States. Population Studies 54, 29–41. Cahill, S., South, K., Spade, J. 2000. Outing age: Public policy issues affecting gay, lesbian, bisexual and transgender elders. Washington, D.C., National Gay and Lesbian Task Force. Clarkberg, M., Stolzenberg, R.M., Waite, L., 1995. Attitudes, values, and entrance into cohabitational versus marital unions. Social Forces 74 (2), 609–634. Fields, J., Clark, C.L. 1999. Unbinding the Ties: Edit Effects of Marital. Status on Same Gender Couples. Population Division Working Paper No. 34, Fertility and Family Statistics Branch. Available at: . Fredriksen-Goldsen, K.I., Muraco, A., 2010. Aging and sexual orientation: A 25-year review of the literature. Research on Aging 32, 372–413. Gates, Gary, 2009. Same-Sex Spouses and Unmarried Partners in the American Community Survey. The Williams Institute, UCLA, Los Angeles, CA. Gates, G.J., 2010. Same-sex couples in US Census Bureau data: Who gets counted and why. Los Angeles, CA: The Williams Institute. . Gates, G.J., Ost, J., 2004. The Gay and Lesbian Atlas. The Urban Institute, Washington, DC. Gary Gates, Steinberger Michael. 2009. Same-sex unmarried partner couples in the american community survey: the role of misreporting, miscoding and misallocation. In: Working Paper Presented at the Population Association of America Annual Meeting, Detroit, MI, April. Grossbard-Shechtman, Shoshana, 1993. On the Economics of Marriage: A Theory of Marriage, Labor, and Divorce. Westview Press, Boulder, CO. Grossman, A.H., 2006. Physical and mental health of older lesbian, gay, and bisexual adults. In: Kimmel, D., Rose, T., David, S. (Eds.), Lesbian, gay, bisexual and transgender aging: Research and clinical perspectives. Columbia University Press, New York, pp. 53–69. Grossman, A.H., D’Augelli, A.R., Hershberger, S.L., 2000. Social support networks of lesbian, gay, and bisexual adults 60 years of age and older. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences 3 (3), 171. Herdt, G.H., de Vries, B., 2004. Gay and Lesbian Aging: Research and Future Directions. Springer, New York, NY. Herek, G.M., 2006. Legal recognition of same-sex relationships in the united states: a social science perspective. American Psychologist 61 (6), 607–621. Horwitz, A.V., White, H.R., 1998. The relationship of cohabitation and mental health: a study of a young adult cohort. Journal of Marriage and the Family 60, 505–514. Hull, Kathleen, 2003. The cultural power of law and the cultural enactment of legality: the case of same-sex marriage. Law & Social Inquiry 28 (3), 629–657. Institute of Medicine (IOM). 2011. The Health of Lesbian, Gay, Bisexual, and Transgender People. Washington, D.C., The National Academies Press. Kiernan, K., 2000. European Perspectives on Union Formation. In: Waite, L.J. (Ed.), The Ties that Bind: Perspectives on Marriage and Cohabitation. Aldine Gruyter, New York, NY. Kiernan, K., 2004. Cohabitation and parenthood in Britain and Europe. Law and Policy 26, 33–55. Kurdek, Lawrence, 2004. Are gay and lesbian cohabiting couples really different from heterosexual married couples. Journal of Marriage and Family 66, 880– 900. Lannutti, Pamela, 2007. The influence of same-sex marriage on the understanding of same-sex relationships. Journal of Homosexuality 53 (3).

A.K. Baumle / Social Science Research 43 (2014) 60–73

73

Lauderdale Diane. 2001. Education and survival: birth cohort, period, and age effects. Demography 38, 551–561. Lupton, J., Smith, J.P. 2003. Marriage, assets and savings. In: Grossbard-Shecht, S. (Ed.), Marriage and the Economy: Theory and Evidence from Advanced Industrial Societies. Cambridge University Press, NY, pp. 129–152. Masini, B.E., Barrett, H.A., 2008. Social support as a predictor of psychological and physical well-being and lifestyle in lesbian, gay, and bisexual adults aged 50 and over. Journal of Gay & Lesbian Social Services 20 (1), 91–110. MetLife. 2010. Still Out, Still Aging: The MetLife Study of Lesbian, Gay, Bisexual, and Transgender Baby Boomers. Westport, CT: MetLife Mature Market Institute. Mirowsky, John, Ross, Catherine E., 1999. Economic hardship across the life course. American Sociological Review 64, 548–569. Musick, K., Bumpass, L., 2006. Cohabitation, Marriage, and Trajectories in Well-being and Relationships. California Center for Population Research, University of California, Los Angeles, CA. O’Connell, M., Gooding, G. 2007. Editing unmarried couples in Census Bureau data. Housing and Household Economic Statistics Division Working Paper. Washington, DC, U.S. Census Bureau. O’Connell Martin, Sarah Feliz. 2011. Same-Sex Couple Household Statistics from the 2010 Census. In: SEHSD Working Paper Number 2011-26. Released September 27, 2011. Peplau, L.A., Beals, K.P., 2004. The family lives of lesbians and gay men. In: Vangelisti, A. (Ed.), Handbook of Family Communication. Erlbaum, Mahwah, NJ, pp. 233–248. Peplau, L.A., Fingerhut, A., 2004. The paradox of the lesbian worker. Journal of Social Issues 60 (4), 719–735. Porche, M.V., Purvin, D.M., 2008. Never in our lifetime: Legal marriage for same-sex couples in long-term relationships. Family Relations 57, 144–159. Riggle, E.D., Rostosky, S.S., Horne, S.G., 2010. Psychological distress, well-being, and legal recognition in same-sex couple relationships. Journal of Family Psychology 24 (1), 82–86. Ross, Catherine E., Mirowsky, John, Goldsteen, Karen, 1990. The impact of the family on health: decade in review. Journal of Marriage and the Family 52, 1061. Rothblum, Esther, 2005. Same-sex marriage and legalized relationships: i do, or do i. Journal of GLBT Family Studies 1, 21–31. Scherrer, K.S., 2009. Images of sexuality and aging in gerontological literature. Sexuality. Research & Social Policy: A Journal of the NSRC 6 (4), 5–12. Seltzer, Judith, 2004. Cohabitation in the United States and Britain: demography, kinship, and the future. Journal of Marriage and Family 66 (4), 921–928. Simmons, T., O’Connell, M. 2003. Married-couple and unmarried-partner households: 2000. Washington, DC, United States Census Bureau, Retrieved May 25, 2006. Available from: . Smock, P.J., 2000. Cohabitation in the United States: An Appraisal of Research Themes, Findings, and Implications. Annual Review of Sociology 26, 1–20. Solomon, S.E., Rothblum, E.D., Balsam, K.F., 2004. Pioneers in partnership: Lesbian and gay male couples in civil unions compared with those not in civil unions and married heterosexual siblings. Journal of Family Psychology 18, 275–286. Solomon, S.E., Rothblum, E.D., Balsam, K.F., 2005. Money, housework, sex, and conflict: same-sex couples in civil unions, and heterosexual married siblings. Sex Roles 52, 561–575. Soons, Judith P.M., Kalmijn, Matthius, 2009. Is marriage more than cohabitation? Well-being differences in 30 European countries. Journal of Marriage and Family 71 (5). South, Scott J., Tolnay, Stewart E., 1992. Relative well-being among children and the elderly: the effects of age group size and family structure. The Sociological Quarterly 33 (1), 115–133. Uhlenberg, Peter (Ed.), 2009. International Handbook of Population Aging. Springer Press, Dordrecht, The Netherlands. Umberson, Debra, 1987. Family status and health behaviors: social control as a dimension of social integration. Journal of Health and Social Behavior 28, 306–319. Umberson, Debra, 1992. Gender, marital status and the social control of health behavior. Social Science and Medicine 34 (1992), 907–917. United States v. Windsor, 570 U.S. ___ (2013). Waite, L.J., 2000. Trends in Men’s and Women’s Well-being in Marriage. In: The Ties that Bind: Perspectives on Marriage and Cohabitation. Aldine Gruyter, New York, NY. Waite Linda. 2005. Marriage and Family. In: Poston Jr., D.L., Micklin Michael (Eds.), Chapter 3 in Handbook of Population. Springer, New York, NY. Waite, Linda J., Gallagher, Maggie., 2000. The Case for Marriage: Why Married People Are Happier, Healthier and Better off Financially. Broadway Books, Westminster, MD. Walther, C.S., Dudley Jr., L.P., 2004. Patterns of gay and lesbian partnering in the larger metropolitan areas of the United States. Journal of Sex Research 41, 201–214. Wight, R.G., LeBlanc, Allen J., Lee Badgett, M.V. 2012. Same-sex legal marriage and psychological well-being: Findings from the california health interview survey. American Journal of Public Health, Published, online. December 13, 2012, e1–e8. .