Women's Experiences of Paid Work and Planning for Retirement 

This section aims to identify factors that are associated with changes in labour force attachment and retirement status over the four surveys between 1996 and 2004. On the assumption that there is particular policy interest in understanding why some women 'prematurely' leave the workforce and, conversely, why others remain in paid work in their mid-age years, particular attention has been paid in the following analyses to four groups: women who have retired 'early', those who have consistently been in paid work over the course of the ALSWH, those who intermittently moved into and out of employment during this period, and the 'late starters' who entered the labour force at some time after 1996 and were still in paid work in 2004.

Methods

In 1996, 13716 mid-age women responded to M1 and 9458 of these women provided answers to questions on paid work at the three subsequent surveys. Women who have dropped out of the ALSWH are likely to have less education (Lee et al., 2005). Also, the women who answered all surveys were more likely to be in paid work at M1 (69%) than those who only answered some of the surveys (59%). Hence the results are likely to overestimate the proportion of women who were in paid work at subsequent surveys.

For the analyses in this section, women were initially placed in one of two groups depending on their paid work status: paid work (PW) or not in paid work (NPW). Women were considered in paid work if they were working full-time, part-time or casual. They were not in paid work if they were doing the following: work without pay, looking for work (unemployed), home duties, unpaid voluntary work, or studying and not involved in paid work. The groups are mutually exclusive. There were four time points when this question was asked, therefore there were 16 possible outcomes.

The 16 outcomes were reduced to 5 categories:

• Always Working: women who were in paid work at all four survey points (n=4432)
• Later starters:women who were not in paid work at M1, but who subsequently joined and stayed in the workforce at subsequent surveys (n=832)
• Early Retirees: women who were in paid work at M1, but left the workforce at some time thereafter and did not return to paid work in subsequent surveys (n=1476)
• Intermittent workers: women who had an erratic pattern of paid work over the four surveys (n=1242)
• Non-workers: women who were not in paid work at any survey (n=1476)

Here we are interested in three models. The first model compares the profiles of women who were early retirees with those who have continued to work over the four surveys (Always Working). The second model compares the profiles of women whose work pattern was erratic with those who were always working. The third model investigates differences in the profiles of women who were not working at M1 but subsequently joined the workforce (Late Starters) with those who were not in paid work over the eight years to 2004.

In the results shown in this section, we will firstly describe the characteristics of the women who were working at all surveys, the reference category for the first two models. The data were analysed using logistic regression models for women who were early retirees (Model 1) or intermittent workers (Model 2) with women who were always working as the reference category. For the third model, we will describe the characteristics of the women who were not working at any survey. Logistic regression was used to analyse differences in predictors for women who were late starters in the workforce compared to women who were non-workers at all surveys.

All models initially included socio-demographic, work and health variables. Educational qualifications (M1) and ability to manage on available income were included to adjust for socioeconomic status. Family and social factors included marital status; dependent children and the provision of care; whether the woman's partner or spouse had been made redundant or retired in the last year; or whether the partner's health had declined in the last year. Work related factors included satisfaction with what the woman had achieved in terms of work and career in her life, and feeling rushed, pressured or busy. The number of general practitioner visits in the last year was used as an objective measure of a woman's health.

Unless otherwise indicated all factors were measured at M4. These factors may be either a cause or an effect of a particular work pattern.

All analyses were adjusted for area of residence to allow for the over-sampling of women living in rural and remote areas. Only those variables that were found to be significantly associated with employment patterns are shown in the tabulated results of the models.

Characteristics of women in paid work at all surveys

Table 10 shows the socio-demographic, family and social, work-related, and health factors that were included in the preliminary analysis of data for mid-age women in paid work at all surveys. Percentages were weighted for area of residence.

The findings for mid-age women in paid work at all surveys were:

• Many women (47%) had completed post-school qualifications by M1
• Most women were married (70%) or living in a de facto relationship (7%); one in eight women was divorced and 4% separated
• More than two-thirds of women found managing on their available income was either not too bad or easy, almost a quarter had difficulty some of the time and 6% found it difficult all the time or impossible
• More than 40% of women still had dependent children either living with them or away from home
• More than a quarter regularly provided care for someone
• Two percent had a spouse or partner who had been made redundant in the last year
• Four percent had a spouse or partner who had retired in the last year
• Seven percent had a spouse or partner who had a major decline in health in the last year
• Most women (92%) were satisfied with their work achievements in their life so far
• Most women (84%) were satisfied with what they had achieved in their career
• Almost two-thirds of women felt rushed, pressured or busy at least a few times a week, 17% felt that way about once a week and 17% rarely felt rushed
• Around 40% of women had up to two general practitioner visits in the last year, 31% had three or four visits and a quarter had five or more visits in the last year

[ top ]

Data are for 4432 women who were in paid work at all surveys

Percentages weighted for area of residence

Three models of work and retirement

Early retirees

Table 11 shows the odds of being an early retiree compared with the odds of always being in work for various socio-demographic, work-related and health variables. For example, the estimated odds ratio of 1.90 in Table 11 (relating to a spouse retiring from work) implies the odds of being an early retiree compared with the odds of always being in work, is about two times higher for women whose spouse had retired compared with those whose spouse had not retired in the last year. The odds ratios do not imply either temporal or causal relationships.

When all the variables were in the model, several factors were not significantly associated with early retirement. These factors were a spouse being made redundant or having a major decline in health and satisfaction with career achievements. While these effects were significant in univariate analyses, once other income, family and work-related factors were included in the model, they were no longer significant.

After adjusting for area of residence, the findings for mid-age female early retirees (compared with those women Always Working) were that:

• Women who had more education had decreased odds of being early retirees
• Women who were not married had decreased odds of being early retirees
• Women who had more difficulty managing on their available income had increased odds of being early retirees, while those who had least difficulty managing on their available income had decreased odds of being early retirees
• Women who had dependent children had decreased odds of being early retirees
• Women who were providing care for someone had increased odds of being early retirees
• Women whose spouse had retired in the last year had increased odds of being early retirees
• Women who were satisfied with their work achievements had decreased odds of being early retirees
• Women who felt rushed most days of the week had decreased odds of being early retirees, while those who rarely felt rushed had more than twice the odds of being early retirees
• Women who had more general practitioner visits had increased odds of being early retirees

[ top ]

Intermittent workers

Table 12 shows the odds of being intermittently in paid work compared with always being in work, for various socio-demographic, work-related and health variables. For example, the estimated odds ratio of 1.70 in Table 12 (relating to difficulty managing on income) implies that the odds of being an intermittent worker compared with the odds of always being in work are almost twice as high for women who find it impossible or always difficult to manage on their income, compared with those who don't have difficulty managing financially. The odds ratios do not imply either temporal or causal relationships.

When all the variables were in the model, several factors were not significantly associated with an erratic pattern of paid work. These factors were having dependent children, a spouse being made redundant, retiring from work or having a major decline in health, and personal satisfaction with achievements in career. While these effects were significant in univariate analyses, once other income, health and work-related factors were included in the model, they were no longer significant.

The findings for mid-age women with an erratic pattern of paid work (compared with those women Always Working) were that:

• Women with more education had decreased odds of being an intermittent worker
• Women who were not married tended to have decreased odds of being an intermittent worker, but this was only significant for divorcees and women who had never married
• Women who had difficulty managing on their available income had twice the odds of being an intermittent worker
• Women who were providing care for someone had increased odds of being an intermittent worker
• Women who were satisfied with their work achievements had decreased odds of being an intermittent worker
• Women who felt rushed most days of the week had decreased odds of being an intermittent worker, while those who rarely felt rushed had more than twice the odds of being an intermittent worker
• Women who had more general practitioner visits had increased odds of being an intermittent worker
• Rural women, and particularly those living in remote areas, had increased odds of being intermittently in paid work

[ top ]

Characteristics of women who were not in paid work at any survey

Table 13 shows the socio-demographic, family and social, work-related and health factors that were included in the preliminary analysis of data for mid-age women who were not in paid work at any survey. Percentages were weighted for area of residence.

The findings for mid-age women who were not in paid work at any survey were:

• Many women (63%) had 10 or fewer years of schooling by M1
• Most women were married (81%), divorced (7%) or widowed (5%)
• More than half the women found managing on their available income was either not too bad or easy, more than a quarter had difficulty some of the time and 18% found it difficult all the time or impossible
• More than 40% of women still had dependent children either living with them or away from home
• Over a third regularly provided care for someone
• Two percent had a spouse or partner who had been made redundant in the last year
• Four percent had a spouse or partner who had retired in the last year
• Ten percent had a spouse or partner whose health had a major decline in the last year
• Most women (79%) were satisfied with their work achievements in their life so far
• Two-thirds of women were satisfied with what they had achieved in their career
• Forty percent of women felt rushed, pressured or busy at least a few times a week, 20% felt that way about once a week and 40% rarely felt rushed.
• One third of women had up to two general practitioner visits in the last year, 27% had three or four visits and 40% had five or more visits in the last year

Data are for 1476 women who were not in paid work at any survey

Percentages weighted for area of residence

Late starters

Table 14 shows the odds of being a late starter compared with not being in paid work at any survey for various socio-demographic, work-related, and health variables. For example, the estimated odds ratio of 2.06 in Table 14 (relating to satisfaction with work achievements) implies that the odds of being a late starter compared with the odds of not being in paid work are twice as high for women who are satisfied with their work achievements, compared with those who are not satisfied with their work achievements in life so far. The odds ratios do not imply either temporal or causal relationships.

When all the variables were in the model, several factors were not significantly associated with being a late starter. These factors were having dependent children, a spouse being made redundant, retiring from work or having a major decline in health, and personal satisfaction with achievements in career. While these effects were significant in univariate analyses, once other income, health, and work-related factors were included in the model, they were no longer significant.

The findings for mid-age women who were 'late starters' in paid work (compared with women who were never in paid employment during the life of the study) were that:

• Women who had more education had increased odds of being a late starter
• Women who were separated had over three times the odds of being a late starter compared to never being in paid work, while those who were divorced or in a de facto relationship had more than twice the odds of making a later entry into the workforce
• Women who had more difficulty managing on their available income had reduced odds of being a late starter
• Women who were providing care for someone had decreased odds of being a late starter
• Women who were satisfied with their work achievements had greater odds of being a late starter
• Women who experienced regular time pressure had increased odds of being a late starter
• Women who had more general practitioner visits had decreased odds of making a later entry into the workforce