Retirement and mental health: Analysis of the Australian national survey of mental health and well-being

Abstract

Nation-wide research on mental health problems amongst men and women during the transition from employment to retirement is limited. This study sought to explore the relationship between retirement and mental health across older adulthood, whilst considering age and known risk factors for mental disorders. Data were from the 1997 National Survey of Mental Health and Well-being, a cross-sectional survey of 10,641 Australian adults. The prevalence of depression and anxiety disorders was analysed in the sub-sample of men ($n=1928$) and women ($n=2261$) aged 45–74 years. Mental health was assessed using the Composite International Diagnostic Instrument. Additional measures were used to assess respondents’ physical health, demographic and personal characteristics. The prevalence of common mental disorders diminished across increasing age groups of men and women. Women aged 55–59, 65–69, and 70–74 had significantly lower rates of mental disorders than those aged 45–49. In contrast, only men aged 65–69 and 70–74 demonstrated significantly lower prevalence compared with men aged 45–49. Amongst younger men, retirees were significantly more likely to have a common mental disorder relative to men still in the labour force; however, this was not the case for retired men of, or nearing, the traditional retirement age of 65. Men and women with poor physical health were also more likely to have a diagnosable mental disorder. The findings of this study indicate that, for men, the relationship between retirement and mental health varies with age. The poorer mental health of men who retire early is not explained by usual risk factors. Given current policy changes in many countries to curtail early retirement, these findings highlight the need to consider mental health, and its influencing factors, when encouraging continued employment amongst older adults.

Introduction

Economic projections show that population ageing in western nations will increase pressure on government expenditure for health and social welfare spending, while the declining proportion of the population participating in the labour force will limit growth in revenue (e.g., Commonwealth Department of the Treasury, 2002). Both international (e.g., Duval, 2003) and Australian domestic (e.g., Commonwealth Department of Family and Community Services, 2002) policy responses have focused on strategies to increase rates of labour-force participation in older adults, including encouraging continued work and delayed retirement.

To successfully adapt social policies to meet these changing circumstances, there is a growing need to understand the health and personal factors that underlie older people's decisions to work and retire. The transition to retirement is often accompanied by shifts in income (e.g., Palmore, Fillenbaum, & George, 1984), and social and physical activity (e.g., Wells & Kendig, 1999). Changes in psychological well-being are also documented (e.g., Gall, Evans, & Howard, 1997; Mein, Martikainen, Hemingway, Stansfeld, & Marmot, 2003). Any association between labour-force participation and mental health may be important to consider when developing employment policies and addressing barriers to increased participation for people of, or nearing, the traditional retirement age.

Retirement is a loosely defined construct. Whilst the official retirement age in most countries (including Australia) is still 65 years, the actual pattern of retirement frequently does not conform to this definition (Leibfritz, 2003). For many individuals, retirement is more accurately described as the time when one's main job ceases and receipt of pension commences, irrespective of age (O’Rand, 1990). The availability of government pensions, and other historically driven incentives, has meant that early retirement has become common in many countries (Blöndal & Scarpetta, 1999; Duval, 2003). Older jobseekers may also redefine themselves as retired, rather than unemployed (Jackson & Taylor, 1994). Given these definitional complexities, retirement is often operationalised as absence from the labour force beyond a certain age (typically 45 years). While this definition has limitations, it has been applied internationally to facilitate cross-national comparisons (e.g., Gendell, 1998; Scherer, 2002), and will be adopted for this study.

Existing literature surrounding labour-force status and mental health has found lower levels of psychological well-being in unemployed compared with working adults (for review see Dooley, Fielding, & Levi, 1996; Murphy & Athanasou, 1999; for discussion see Bartley, 1994; Ezzy, 1993). However, such research has generally focused on employment within young and middle age populations. Theory and research regarding the effects of retirement on mental health are inconsistent. Some argue that retirement can relieve a person of the stressors of the working environment, and increase their sense of personal control (Drentea, 2002; Kim & Moen, 2002). Conversely, retirement may have an adverse effect through the absence of positive benefits derived from employment, including: financial stability, a sense of purpose, or social activity (Moen, 1996).

Cross-sectional studies have shown both that retirees are more likely to report poor psychological well-being compared with their employed peers (e.g., Bosse, Aldwin, Levenson, & Ekerdt, 1987) or that there is no association between retirement and measures of mental health or psychological distress (e.g., Herzog, House, & Morgan, 1991; Ross & Drentea, 1998). These discrepancies have not been resolved by longitudinal investigations, which have typically focused on occupational cohorts. While some studies have reported increased depression and/or anxiety symptoms (e.g., Richardson & Kilty, 1991), others have found improved mental health following retirement (Mein et al., 2003). Factors that may moderate the relationship between retirement and mental health include: gender (e.g., Kim & Moen, 2002; Richardson & Kilty, 1995); self-efficacy (e.g., Wells & Kendig, 1999); psychological well-being prior to retirement (e.g., Kim & Moen, 2002); work and social circumstances prior to retirement (e.g., Reitzes, Mutran, & Fernandez, 1996), whether retirement was voluntary or not (e.g., Swan, Dame, & Carmelli, 1991); physical health (e.g., Gall et al., 1997); and socio-economic status (e.g., Mein et al., 2003).

It is important to note that both age and cohort may confound interpretation of research in this area. While age is clearly associated with retirement, it is also a measure of biological and psychological functioning, a determinant of social roles and norms, and an indicator of birth-cohort membership (Moen, 1996). A number of epidemiological studies have demonstrated that older adults experience a lower prevalence of common mental disorders and psychological symptoms than middle-aged men and women (e.g., Andrews, Hall, Teesson, & Henderson, 1999; Henderson et al., 1998; Melzer, Buxton, & Villamil, 2004; though see Mirowsky & Ross, 1992). Therefore, better mental health observed at retirement may reflect age or cohort differences.

The psychological effects of labour-force separation may be influenced by the social norms surrounding employment status at different ages (Herzog et al., 1991). Retirement is the norm for men aged 65 and older in most western nations, with most men below this age still in paid employment. Herzog et al. (1991) proposed that the adverse psychological effects of retirement might be greater for younger retirees relative to those at or above the official retirement age (see also Palmore et al., 1984).

Melzer et al. (2004) reported an inverse association between age and the prevalence of common mental disorders in a cross-sectional sample of older men in the National Psychiatric Morbidity Survey of Great Britain. A strikingly lower prevalence of mental disorders coincided with Britain's official retirement age of 65 years, regardless of workforce status. Young retirees (i.e., those below the age of 65) had poorer mental health than their working peers, but retrospective data showed that men who had retired early and since reached the age of 65 displayed prevalence rates similar to working men. Melzer and colleagues suggested that the association between mental health and retirement age could reflect that labour-force inactivity is considered socially unacceptable for those below the traditional retirement age.

Our aim is to examine the prevalence of common mental disorders (depression and anxiety disorders) in a sample of older Australians, replicating the British findings. We investigate whether there is lower prevalence of common mental disorders coinciding with Australia's traditional retirement age (65 years). We also consider the effect of labour-force status with respect to age differences and the influence of demographic characteristics and risk factors.