Retirement planning is an important midlife task, increasingly so because of greater longevity, uncertainty about pensions and future health care costs (McDonald & Donahue, Reference McDonald and Donahue2011), and ambiguous choices employees sometimes face between defined benefit and defined contribution retirement plans (Lusardi & Mitchell, Reference Lusardi and Mitchell2005). Greater retirement planning enhances confidence in the retirement transition (Taylor-Carter, Cook, & Weinberg, Reference Taylor-Carter, Cook and Weinberg1997) as well as satisfaction with retirement (Quick & Moen, Reference Quick and Moen1998). Researchers are developing an understanding of diversity in retirement planning experiences and how retirement planning may be shaped by gender (Smith & Moen, Reference Smith and Moen1998; Szinovacz, DeViney, & Davey, Reference Szinovacz, DeViney and Davey2001), race, or ethnicity (Richardson & Kilty, Reference Richardson and Kilty1992; Silverman, Skirboll, & Payne, Reference Silverman, Skirboll and Payne1996), and a number of studies carried out over the past decade have examined how sexual minority status (e.g., gay, lesbian, bisexual) has an impact on retirement planning (Metlife Mature Market Institute & The Lesbian and Gay Aging Issues Network of the American Society on Aging, 2010; Mock & Cornelius, Reference Mock and Cornelius2007; Mock, Sedlovskaya, & Purdie-Vaughns, Reference Mock, Sedlovskaya and Purdie-Vaughns2011).
One particularly important challenge for many sexual minorities is homophobia and the impact – known as minority stress – it has on their well-being (Mays & Cochran, Reference Mays and Cochran2001; Meyer, Reference Meyer2003). A defining characteristic of minority stress is the sense of social exclusion or lack of support created by homophobia. Given the importance of social support for future orientation and planning (Emmons, Reference Emmons1986; Hershey & Mowen, Reference Hershey and Mowen2000), perceived social support may be a particularly important factor for retirement planning among sexual minorities. Drawing on theory related to marginalized identities and those with low power, we propose that sexual minorities may be particularly mindful of social cues and information (Frable, Blackstone, & Scherbaum, Reference Frable, Blackstone and Scherbaum1990; Keltner, Gruenfeld, & Anderson, Reference Keltner, Gruenfeld and Anderson2003). Thus, in the present study, we examined the retirement planning perceptions of sexual minority adults focusing on the role of perceived social support.
Retirement Planning
Retirement has traditionally been defined as a late-life withdrawal from the workforce, but this definition masks the increasing complexity in the nature of that transition. For example, some people may retire but then return to work because of financial needs or because they missed the social contacts and challenge of the workplace (Schellenberg, Turcotte, & Ram, Reference Schellenberg, Turcotte and Ram2005). Retirement planning is a complex process that involves planning for health care needs, housing arrangements, leisure in retirement, and potential post-retirement employment or volunteering (Moen, Reference Moen, Mortimer and Shanahan2003). Financial security is a key factor required to carry out most other retirement activities (Dennis & Migliaccio Reference Dennis and Migliaccio1997), underscoring the importance of retirement financial planning. Typical factors that shape retirement planning include age, income, and family composition. To be specific, people tend to plan more the older they are (Kosloski, Ekerdt, & DeViney, Reference Kosloski, Ekerdt and DeViney2001) and the greater their incomes (Phua & McNally, Reference Phua and McNally2008). Those with children tend to plan less for retirement, and unmarried men are less likely to retire than other men and women (Szinovacz et al., Reference Szinovacz, DeViney and Davey2001).
Research on retirement timing (as opposed to intentions or planning) sheds light on pre-retirement circumstances that have an impact on retirement planning. For example, health concerns have been found to be an even more important factor in the retirement decision than income security and employment satisfaction (Cobb-Clark & Stillman, Reference Cobb-Clark and Stillman2006; Dwyer & Mitchell, Reference Dwyer and Mitchell1999). Dwyer and Mitchell (Reference Dwyer and Mitchell1999) noted that men who reported poor health expected to retire up to two years sooner than did their healthy counterparts. Similarly, longitudinal research with Canadian data shows that poor health hastens the retirement transition as does pension wealth (Shirle, Reference Shirle2010). In addition to considerations of timing, diverse demographic characteristics also shape retirement financial planning. A recent cross-cultural study has highlighted the contribution of financial knowledge as well as financial resources (e.g., income, employer pension, and assets) to financial planning activities and perceptions of adequacy of retirement savings (Hershey, Henkens, & van Dalen, Reference Hershey, Henkens and van Dalen2010). In sum, multiple demographic characteristics – including gender, wealth, family composition, health, and access to pensions – shape retirement and the retirement planning process.
Psychological Factors Associated with Planning
In addition to these demographic and structural factors, psychological resources and perceived support have been linked to future orientation and planning in general but also to retirement planning in particular. The pursuit of goals and the planning that accompanies this process gives structure and meaning to life but is also linked to challenges, setbacks, and accomplishments that both undermine and enhance well-being (Brandtstädter & Rothermund, Reference Brandtstädter and Rothermund2002). Conversely, well-being is also an antecedent of planning, with distress and diminished well-being having a negative impact on goal-directed behaviour (Emmons, Reference Emmons1986). To be specific, anxiety has been linked to a greater likelihood of pursuing avoidance goals (e.g., avoiding negative situations) of depression, and of decreased involvement in approach goals (e.g., achievement-focused goals) (Dickson & MacLeod, Reference Dickson and MacLeod2004). These broader findings are reflected in more specific research on psychological factors and how they relate to retirement planning. For example, emotional stability (e.g., emotional lability, moodiness) and future time perspective (e.g., enjoyment from contemplating the future, future goal orientation) are both linked to retirement planning (Hershey & Mowen, Reference Hershey and Mowen2000).
Perceived Support, Well-being, and Planning
Social support plays a critical role at the intersection of well-being, future orientation, and planning. Social ties and perceived social support can be both intrinsically rewarding and also fundamental for well-being by providing emotional and tangible resources (e.g., Cohen & Wills, Reference Cohen and Wills1985; Uchino, Cacioppo, & Kiecolt-Glaser, Reference Uchino, Cacioppo and Kiecolt-Glaser1996). In fact, one of the ways social resources enhance well-being is through social support that enhances personal goal striving (Diener & Fujita, Reference Diener and Fujita1995). Other research on future orientation and planning has found that both psychological well-being (e.g., lower neuroticism) and perceived support (e.g., presence of emotionally supportive and caring relationships) are linked to greater planning (Prenda & Lachman, Reference Prenda and Lachman2001). Thus, perceived support likely enhances future orientation and planning through support for goal striving (e.g., Diener & Fujita, Reference Diener and Fujita1995) as well as by the role that support plays in enhancing psychological well-being.
Sexual Minority Status and Perceived Support
A consideration of the role of social support in future orientation and perceptions of retirement planning touches on perhaps one of the greatest challenges for the well-being and retirement planning of sexual minority adults. To be specific, despite recent advances towards equality for sexual minorities in Canada (Nicol & Smith, Reference Nicol and Smith2008), homophobic stigmatization persists (Morrison, Morrison, & Franklin, Reference Morrison, Morrison and Franklin2009; Rye & Meaney, Reference Rye and Meaney2009). Both the direct experience and potential threat of stigmatization create considerable distress for many sexual minority individuals (Mays & Cochran, Reference Mays and Cochran2001; Meyer, Reference Meyer2003). This distress, characterized as minority stress (Meyer, Reference Meyer2003), stems largely from internalization of negative social cues that devalue sexual minority identities (i.e., internalized homophobia; Herek, Cogan, Gillis, & Glunt, Reference Herek, Cogan, Gillis and Glunt1997) and from a sense of alienation and social rejection caused by stigmatization (Durkheim, Reference Durkheim1951; Goffman, Reference Goffman1963; Meyer, Reference Meyer2003).
In addition to the potentially negative impact of the minority stress process for sexual minorities, marginalization has also been linked to heightened vigilance during social interactions. Research by Frable et al. (Reference Frable, Blackstone and Scherbaum1990) shows that in social interactions, those with non-normative identities and, in particular, concealable identities (e.g., sexual minorities), had more detailed recall of their surroundings during social interactions than those with normative identities; engaged in more perspective-taking; and paid closer attention to partners’ behaviours (Frable et al., Reference Frable, Blackstone and Scherbaum1990). Those with a concealable stigma could also be said to have relatively lesser power in social situations. Research examining the role of power in perceptions and attitudes shows that those with lower (vs. higher) social power are more vigilant to social cues and information (e.g., scanning for potential threat and judgement by others) (Keltner et al., Reference Keltner, Gruenfeld and Anderson2003). This heightened vigilance to social cues, coupled with the minority stress process outlined above, suggests that perceived social support, although a potentially precarious resource for sexual minorities, may also be something to which they are keenly attuned.
Summary and Hypotheses
In the present study, we examined the role of perceived social support in the retirement planning perceptions of sexual minorities. Previous research on planning in general and retirement planning in particular suggests that perceived support is associated with enhanced planning (Hershey & Mowen, Reference Hershey and Mowen2000; Prenda & Lachman, Reference Prenda and Lachman2001). For sexual minorities, stigmatization and rejection (e.g., Meyer, Reference Meyer2003) may not only disrupt typical social support processes but also make them vigilant to social information such as perceived support (Frable et al., Reference Frable, Blackstone and Scherbaum1990; Keltner et al., Reference Keltner, Gruenfeld and Anderson2003). Thus, we expect that although greater perceived support may be associated with enhanced retirement planning perceptions in general, perceived support will have a particularly strong impact on the retirement planning perceptions of sexual minorities.
Methods
Data were drawn from the nationally representative General Social Survey (GSS), Cycle 21 (Statistics Canada, 2007). In the GSS 21 survey, more than 23,000 adults aged 45 and older provided responses to questions focused on sexual identity, social support, retirement planning, and other measures (e.g., caregiving) not included in the current study. Data were collected using random-digit dialing with a response rate of 57.7 per cent, and interviews were approximately 40 to 45 minutes in length. To focus on adults for whom retirement planning is likely a particularly salient issue, analyses were conducted with respondents between the ages of 45 and 70, who identified their sexual orientation, and who were not retired but reported that they did plan to retire eventually, yielding a final sample of 6,068 participants.
Measures
Demographics
Age was calculated in years based on reported date of birth. Response options for gender were coded as male = 1 or female = 2. Trans status was not assessed. Marital status options were re-coded for the present analyses as a series of dichotomous variables including cohabiting, widowed, separated or divorced, and single (never married) with married as the reference group. Participants’ children were taken into account with two measures, one that assessed presence of children at home (yes = 1, no = 0), and the other assessing children out of home (yes = 1, no = 0). Education was assessed at five levels (1 = some secondary/elementary/no schooling, 2 = high school diploma, 3 = some university/college, 4 = diploma or certificate from college, 5 = university degree, bachelor’s or greater). Employed was defined by participant responses to a measure in which they identified their main activity during the past 12 months as working at a paid job or business (1), or some other activity (0). Other activities included looking for paid work, going to school, household work, and so on. Income was measured ranging from no personal income = 1, to less than $5,000 = 2, then to increments up to $80,000 to $99,000 = 11 and $100,000 and more = 12. Workplace pension was assessed with a question asking participants if they had a pension plan through current employment (yes = 1, no = 0). Participants’ ratings of physical health were reverse-scored to 1 = poor, 2 = fair, 3 = good, 4 = very good, and 5 = excellent. Similarly, self-rated mental health was recoded on the same scale.
Perceived support was calculated as the mean of three items (i.e., “There are plenty of people that I can rely on when I have problems,” “There are many people I can trust completely,” and “There are enough people I feel close to”), answered with response options from no (1) to more or less (2) or yes (3) (α = .72; De Jong Gierveld & Van Tilburg, Reference De Jong Gierveld and Van Tilburg2006).
Participants were asked to describe their sexual orientation identities with the following options: heterosexual, bisexual, homosexual (gay or lesbian), or unknown. For the purposes of the analyses, two sexual minority variables were constructed, one for gay/lesbian (1) with heterosexual as the reference (0) and bisexual (1), again, with heterosexual as the reference (0). Recall that for the present analyses, we included only participants with complete information for sexual orientation.
Retirement Planning
We assessed retirement planning perceptions with three different items. For Expected Retirement Age, participants were asked to report, in years, the age at which they planned to retire. Participants rated their Certainty of Retirement Age on a scale from 1 (“not at all certain”) to 2 (“somewhat certain”) to 3 (“very certain”). Finally, respondents rated their Anticipated Income Adequacy in retirement on a 5-point scale coded from 1 (“very inadequate”) to 5 (“more than adequate”).
Analysis Plan
Analyses began with the calculation of means, standard deviations, and frequencies for all variables. We used linear regression analyses to test the association of the three retirement planning criteria variables (expected retirement age, certainty of retirement age, and anticipated income adequacy in retirement), with demographic characteristics, a measure of social support, and respondents’ self-identified sexual orientation identity. Two regression models were analysed for each criterion variable with demographic factors, social support, and sexual minority status included in Model 1. Sexual minority status included bisexual and gay/lesbian identity with heterosexual as the reference group. To test the hypothesis that perceived social support may have a stronger association with retirement planning perceptions for sexual minorities compared to heterosexuals, Model 2 included all variables from Model 1 with the addition of sexual minority status (e.g., bisexual or gay/lesbian) by perceived support interaction terms.
Any statistically significant interaction terms were probed by calculating simple slopes at relatively low or high levels of the moderator using Hayes’ Process macros for SPSS (Hayes, Reference Hayes2013). For the current analyses, simple slopes represent the association of sexual minority status (vs. heterosexual) with each criterion variable at low (M – 1 SD), and high (M + 1 SD) levels of the moderator variable (i.e., perceived support). Finally, the GSS sampling weight was adjusted to account for the selection criteria and applied to univariate and multivariate analyses in order to account for the number of people represented by each respondent within the population (Statistics Canada, 2007).
Results
On average, participants were approximately 53 years old, just over half were women, and approximately 67 per cent were married (Table 1). Just under half reported having children living in the home and just under half reported having children who were not currently in the home. Sixty-four participants (0.7%) reported a bisexual identity and 101 (1.1%) reported a gay or lesbian sexual orientation identity with the remainder reporting a heterosexual identity. The average level of education was 3.32, slightly above the mid-point of “some university or college”. Just over 80 per cent were employed, and the average annual personal income level of 7.68 corresponded approximately to $40,000 to $49,999 per year. Forty-two percent reported having a workplace pension.
Table 1: Means and percentages of demographics, socioeconomic status, well-being, and retirement expectations
Mean ratings for physical health and for mental health were both above the mid-point of “good”. Similarly, the mean value for social support was above the mid-point of “more or less” (Table 1). Finally, the average expected retirement age was just over 61, the mean rating for certainty of retirement age was just over the mid-point of “somewhat certain”, and the average rating for anticipated income adequacy was over the mid-point of “barely adequate”.
Regression Results
Expected Retirement Age
Expected retirement age was later for older participants and those who were divorced or single (vs. married), currently working for pay (compared to those who were not), and those with a higher level of education (Table 2, Model 1; Retirement Age). In contrast, expected retirement age was lower for women compared to men, and was also lower the more income participants earned and the more supported they felt (Table 2, Model 1; Expected Retirement Age). Sexual minority status (e.g., bisexual or gay/lesbian vs. heterosexual) was not associated with expected retirement age. However, the gay/lesbian by perceived support interaction term was statistically significant (Table 2, Model 2; Expected Retirement Age). To better understand the nature of the interaction, simple slopes were calculated and showed that at lower levels of perceived support, the gay/lesbian respondents had an older expected retirement age than did heterosexual participants (b = 1.52, SE = 0.70, p < .05). However, at higher levels of perceived support, those who identified as gay or lesbian did not differ from those who identified as heterosexual in terms of expected retirement age (b = –0.53, SE = 0.55, p = .43).
Table 2: Regression coefficients showing the association of retirement expectation criterion variables with demographics, social support, sexual minority status, and sexual minority status by perceived social support interaction terms
* p < .05, ** p < .01, *** p < .001
a male is reference group; b married is reference group; c heterosexual is reference group; Model 2 includes all Model 1 variables
Certainty of Retirement Age
Certainty of retirement age was greater the older the participants were, the more income they earned, and for those with a workplace pension (vs. those without a workplace pension) (Table 2, Model 1; Certainty of Retirement Age). Similarly, those with better self-rated physical health and mental health, and also higher levels of perceived support, reported greater certainty about expected retirement age (Table 2, Model 1; Certainty of Retirement Age). Uncertainty about retirement age was greater for participants in cohabiting relationships, separated/divorced, or single (vs. married), participants with any children in the home, and those who were currently employed (Table 2, Model 1; Certainty of Retirement Age). Sexual minority status was not statistically associated with certainty of retirement age. The gay/lesbian (vs. heterosexual) by perceived support interaction term was statistically significant (Table 2, Model 2; Certainty of Retirement Age). Probing this interaction showed that at lower levels of support, gay or lesbian participants were less certain about their retirement age than heterosexual participants (b = –0.25, SE = 0.11, p < .05). At higher levels of perceived support, sexual minority status was unrelated to certainty about age of retirement (b = 0.11, SE =.09, p = .20).
Anticipated Income Adequacy
Anticipated income adequacy in retirement was greater for older respondents, women (compared to men), those with a greater personal income, workplace pension, better physical health and mental health, and those with higher levels of perceived social support (Table 2, Model 1; Anticipated Income Adequacy). Conversely, individuals who were currently cohabiting, divorced or single (compared to those who were married), with children in or out of the home, or employed, reported lower levels of anticipated retirement income adequacy (Table 2, Model 1; Anticipated Income Adequacy). Sexual minority status was not associated with anticipated income adequacy. The gay/lesbian (vs. heterosexual) by perceived support interaction term was statistically significant (Table 2, Model 2; Anticipated Income Adequacy). Probing this interaction term showed that at lower levels of perceived support, anticipated adequacy of retirement income did not differ significantly by sexual minority status (b = –0.14, SE = 0.12, p = .35). Similarly, at higher levels of perceived support, the simple slope was not statistically significant (b = 0.10, SE = 0.09, p = .28). To clarify, although higher or lower support perceptions did not have a statistically significant association with sexual minorities’ retirement planning, these simple slopes were significantly different from each other and represent a spreading interaction. To be specific, there is a trend that lower perceived support is linked to lower anticipated income adequacy for gays and lesbians (compared to heterosexuals), and greater perceived support is linked to higher levels of anticipated income adequacy for gays and lesbians (compared to heterosexuals).
Discussion
With the present study, we sought to compare the retirement planning perceptions of sexual minority and heterosexual adults drawing on nationally representative data and to examine the potential role of perceived support in shaping retirement planning perceptions. Although our focus was on sexual orientation identity and perceived support, some consistent patterns emerged in the regression analyses. For example, being separated or divorced as well as being single (vs. married) was associated with later expected retirement age, and less certainty regarding retirement age and income adequacy in retirement. Conversely, greater income and having a workplace pension was associated with younger expected retirement age and greater certainty regarding age and income adequacy. The more supported that participants felt, the younger their expected retirement age and the greater their certainty about retirement age and anticipated income adequacy. Finally, sexual minority adults did not differ significantly from heterosexual adults in terms of expected retirement age, certainty of retirement age, or anticipated income adequacy in retirement.
Given the potential challenges sexual minorities face related to stigmatization and the minority stress process outlined above (e.g., Meyer, Reference Meyer2003), as well as the role of psychological resources and perceived support in future orientation and retirement planning (e.g., Hershey & Mowen, Reference Hershey and Mowen2000; Prenda & Lachman, Reference Prenda and Lachman2001), it is perhaps surprising that in a direct comparison, no significant differences were found in terms of sexual orientation identity and retirement planning perceptions. Consistent with previous research on perceived support, future orientation, and planning (e.g., Diener & Fujita, Reference Diener and Fujita1995; Prenda & Lachman, Reference Prenda and Lachman2001), the more supported participants felt (regardless of sexual minority status), the earlier they expected to retire and the more certain they were about their retirement timing and finances.
However, perceived social support appears to play a particularly important role for the retirement planning perceptions of sexual minorities. In particular, for gay and lesbian (but not bisexual) adults, lower perceived support was associated with a later retirement age and less certainty about retirement age, and there was some evidence that lower levels of perceived support were linked to lower anticipated retirement income adequacy for sexual minority adults. This finding is consistent with research that shows those with marginalized identities or with lesser power in social situations – in particular, those with concealable marginalized identities – process social information more thoroughly than those who are not marginalized (Frable et al., Reference Frable, Blackstone and Scherbaum1990; Keltner, et al., Reference Keltner, Gruenfeld and Anderson2003). Specifically, the current findings suggest that this more thorough processing of social support perceptions (whether positive or negative) may explain why perceived support has a greater impact on gay and lesbian adults compared to heterosexual adults. Other social psychological research supports this association: namely, that more socially oriented groups (e.g., people from Eastern vs. Western cultures; those with low vs. high socioeconomic status) are more strongly affected by social information (e.g., Markus & Kitayama, Reference Markus and Kitayama2010).
A few limitations to the current study are worth noting. First, the data are cross-sectional, precluding causal claims. Next, although we predicted that perceived support would be particularly important for sexual minorities resulting from their marginalized status, the data set we drew from did not have information on perceived stigmatization to more directly assess how marginalized participants may feel. Additionally, although our findings highlight the importance of social support for retirement planning perceptions among gay and lesbian adults, we were unable to consider the sources of this support.
The nature and sources of support have been found to differ for sexual minority adults compared to the general population. For example, some research suggests sexual minority adults turn to close friends or families of choice for support rather than biological kin (Oswald, Reference Oswald2002). Lesbian, gay, and bisexual adults are also more satisfied with the support they receive from those who know about their sexual orientation compared to those who do not (Grossman, D’Augelli, & Hershberger, Reference Grossman, D’Augelli and Hershberger2000). Thus, future research on the impact of social support on development and well-being of sexual minority adults would benefit from a more finely grained consideration of the source and nature of the support. Finally, sexual minority status is sometimes more broadly defined and includes, in addition to sexual orientation identity, a more diverse consideration of gender (e.g., trans) than was available in the data. However, this study does represent one of the first overviews of the nature of retirement planning perceptions across diverse sexual orientation identities with nationally representative data. It is worth noting that although the link of perceived support to retirement planning perceptions was particularly strong for gay and lesbian adults, this was not the case for bisexual adults. We recommend future research be done to explore diversity among sexual minorities in terms of their retirement planning. In sum, this study adds to a growing body of research that shows how sexual minority status shapes retirement planning and highlights the unique role of perceived support for those with a marginalized identity.
