6238.0 - Retirement and Retirement Intentions, Australia, July 2012 to June 2013 Quality Declaration 
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 09/12/2013   
   Page tools: Print Print Page Print all pages in this productPrint All


With the first of the 'baby boomer' generation reaching 65 years of age, retirement intentions and behaviours are an area of increasing interest. Data on retirement intentions can be used to inform on social policy issues, such as the upcoming increase in the age pension eligibility age, and adequacy of government payments to and services for Australia's aging population. The data can also be used to inform on the effectiveness of government taxation and superannuation policies. This article uses a cohort analysis approach to compare the retirement intentions and behaviours of four age cohorts in 2012–13 against the expectations of each of the same cohorts in the past.

Conventional analysis of differences in retirement age and retirement intentions compares different groups of people at the same age (eg 65–69 year olds in 2004–05 compared to 65–69 year olds in 2012–13), or different age groups in the same year (eg 60–64 year olds compared to 65–69 year olds in 2012–13). A limitation of these types of comparison is that the choices of the two groups of people will be impacted differently by social, economic and policy factors.

In contrast, the cohort method enables differences between age cohorts to be observed, independent from the effects of year (time period) and age. In addition, it enables more accurate observation of the impacts of year and age by controlling for the age cohort variable1. The Retirement and Retirement Intentions Survey (R&RI) is not a longitudinal survey2, but compared to traditional cross-sectional analysis, comparing different samples from the same age cohort over time provides a better idea of how retirement intentions and behaviours for the same age cohort have changed between 2004–05 and 2012–13.

The age cohorts used in this article have been structured in five-year groupings based on respondents' ages at the time of 2012–13 interview, and traced backwards to find the equivalent age group at the time of each previous survey. The age ranges represented by each cohort in each survey year are outlined in the table below. The age cohort concept is explained further in Appendix 3: Information about the cohort analysis method.

Age at time of interview (years)






Cohort 147-5149-5351-5553-5755-59
Cohort 252-5654-5856-6058-6260-64
Cohort 357-6159-6361-6563-6765-69
Cohort 462-6664-6866-7068-7270-74

This article considers the relationship between expected and actual behaviour for the following three aspects of retirement and retirement intentions:

  • Timing of retirement;
  • Transition to retirement; and
  • Main source of income in retirement.

Cohort analysis allows a consideration of whether intended retirement age corresponds to actual retirement age, and whether this relationship differs between age cohorts.

In each survey from 2004–05 to 2012–13, the average actual retirement age is lower than the average intended retirement age. Such a comparison contrasts past behaviour of people who have already retired, with anticipated future behaviour of people who have not yet retired. In contrast to this method, the analysis below describes the relationship between intended and actual retirement behaviours of the same group of people.

Figure 1 compares the retirement status of four age cohorts, grouped by the respondents' age in 2012–13. It shows their expected retirement status in 2012–13 from earlier surveys (2004–05, 2006–07, 2008–09 and 2010–11), and actual retirement status in the 2012–13 survey. The expected retirement status in 2012–13 is based on the expected age of retirement reported in each survey.

Figure 1 Expected and actual retirement status

The overall proportion of people expecting in 2004–05 to be retired by 2012–13, was higher than the proportion of people that actually was retired by 2012–13, and the proportion expecting to be retired by 2012–13 generally decreased over time. This may be due to people being better able to estimate their short-term financial, health and social situation. Younger people tend to perceive retirement as a time of leisure, but the proportion of people with that perception decreases with age3. However, differences are evident between the cohorts in relation to this general pattern.

The expectations of cohort 4 (age 70–74 years in 2012–13) did not change significantly over the eight year period. The proportion of people in this cohort who were actually retired in 2012–13 was similar to the proportion of people who expected in 2004–05 that they would be retired by 2012–13.

For cohorts 1, 2 and 3 (age 55–69 years in 2012–13), the percentage of people expecting in 2004–05 to retire within eight years was much higher than the percentage of people who did retire by 2012–13. This shows that people are staying in the labour force longer than the reported intentions of the cohort eight years earlier in 2004–05.

The expectations of cohort 3 (age 65–69 years in 2012–13) became more accurate with time, but the proportion expecting in 2010–11 to be retired within two years was still higher than the proportion that was actually retired in 2012–134. For cohorts 1 and 2, the percentage expecting to be retired by 2012–13 also decreased over time, but for these younger cohorts, the decrease lead to an underestimate of the percentage of people that would be retired by 2012–13. This shows that people in the younger cohorts retired later than the expectations of the cohort eight years before in 2004–05, but retired earlier than the expectations of the cohort four years before in 2008–09 (cohort 2) and two years before in 2010–11 (cohorts 1 and 2).

The change between expectation and actual behaviour can be impacted by economic, social and policy factors. In 2008–09, cohorts 2 and 3 (age 60–69 years in 2012–13) in particular experienced a drop in the proportion of people expecting to be retired by 2012–13, which may be due to the economic uncertainty around that time because of the global financial crisis (GFC). The GFC led to a downturn in investment markets, and was an anxious time for superannuation fund members.

Overall, people are seen to be staying in the labour force longer than cohort expectations indicated six or eight years earlier, and for cohort 3, longer than expectations indicated just two years earlier. This may be in the result of more people transitioning to part-time work rather than retiring, which is explored in the following section.


A change from full-time employment to retirement represents a significant change – not just financially, but in terms of lifestyle, self-image and purpose. To assist in adjusting to this change, people who have previously worked full-time may decide to transition in to retirement by cutting back on hours or level of responsibility at work prior to retiring. With increases in non-standard working arrangements across the general population, such transition to retirement behaviours are becoming more possible.

Changes in part-time work

It was seen above that people are remaining in the workforce longer than the previous expectations of their cohorts indicated, and that the short term (two year) expectations of cohort 3 overestimated retirement. This observation is supported by an increase in the proportion of people working part-time among this cohort. The proportion of workers employed part-time increased between the 2004–05 and 2012–13 surveys, from 30% to 49% for cohort 3 and from 38% to 57% for cohort 4.

The increased proportion of part-time workers reflects a steady decrease in the number of people working full-time, while the number of people working part-time has remained relatively stable over the period (see figure 2). This may mean that full-time workers are retiring earlier while part-time workers are working longer. Alternatively, it may result from people moving from full-time to part-time employment, with retirement occurring at a similar rate for both full-time and part-time workers. It is not possible to determine which of these explanations is correct, as it is not known whether the people working part-time in 2012–13 are the same individuals who were working part-time in 2004–05 (because the survey is not longitudinal).

Figure 2 Number of people working full-time and part-time

Preference for more hours of work

If people are moving from full-time work to part-time work before retirement, the next question is whether this is occurring voluntarily (as people transition into retirement for leisure related reasons) or involuntarily (such as age discrimination causing older workers difficulty in finding full-time employment5). In the latter case, the trend towards part-time employment would constitute an increase in underemployment6, rather than a transition to retirement.

Looking at whether or not people working part-time hours would prefer to work more than their usual hours may help distinguish between these causes. For all cohorts, people usually working 15 hours or less but preferring to usually work more hours accounted for between one and three percent of all employed people across all years. The increase in the proportion of people working 15 hours or less was driven by people who did not prefer to work more than these hours. This is shown in figure 3, for cohorts 3 and 4 combined.

Percentage of all Employed, Cohorts 3–4

Figure 3 People working 0-15 hours by preference

In 2012–13, the proportion of people who prefer more hours was similar for people usually working 0–15 hours and for people working 16–34 hours. Around 4% of all employed people in cohorts 3 and 4 were working part-time (less than 35 hours) but wanted more hours, and 46% were working part-time and did not prefer more hours. 2012–13 was the first time that people working 16–34 hours per week were asked about their preference to work more hours in this survey.

Together, the first two sections of this article have considered the timing of retirement, and indicated that people around traditional retirement age (age 65 years) are:

  • remaining in the labour force for longer than their previous expectations indicated;
  • are more likely to be doing so on the basis of part-time employment; and
  • generally would not prefer to work any more hours than they usually do.


As discussed above, it is useful for policy makers to have a good indication of key aspects of people's future retirement. One key policy focus is people's ability to support themselves financially in retirement, either being self funded (through superannuation, investments or other personal income sources) or by relying on government pensions (such as age or disability pension).

A static (single year) comparison between the actual main source of income of the retired population and the expected main source of income from the non-retired population will be particularly impacted by changes in superannuation laws (such as the 1992 introduction of compulsory employer contributions). Such changes will clearly have a greater impact on those who remain in the labour force for longer. Cohort analysis will help to distinguish between differences in expectations by year, by age and by cohort and, importantly, how well they align with actual retirement experience.

Expectation before retirement

In 2004–05, most people expected that they would be self funded in retirement (64% for cohort 1, falling to 46% for cohort 4), rather than reliant on a government pension or other sources. This is despite government pensions being the most common actual main source of personal income in the same year for people in the same age cohort who had already retired.

Actual main source of income at retirement

For people who retired between 2004–05 and 2012–13, less people were mainly self funded at retirement than expected to be in 2004–05. This was true for every cohort, as seen in figure 4.

Figure 4 Percentage mainly self-funded in retirement

Government pension(s) was the main source of income at the time of retirement for a higher proportion of people than expected this in 2004–05. This was true for all cohorts except cohort 2. Cohort 2 was around the average retirement age in 2012–13 (61.5 years for recent retirees7), and had the highest proportion of self funded retirees (as shown in figure 4), as well as the lowest proportion of people relying on government pensions (figure 5).

Those from cohort 1 who retired in the period between 2004–05 and 2012–13 were well below the average retirement age, and had not had as much time to accumulate superannuation. Most people in cohort 1 who retired during this period and were on a government pension were, in 2012–13, receiving a disability pension (55%), and may have needed to retire earlier than expected because of health circumstances. The experience of those who have retired at a particularly young age therefore may not be representative of the cohort as a whole.

Cohort 3 had the biggest difference between expected source of income and actual source of income at retirement. This can be seen in figure 5, along with actual source of income after retirement.

Actual main source of income after retirement

A higher reliance on government pension at retirement compared to expectation (apart from cohort 2) was discussed above. In addition, there is an increasing reliance on a government pension between the beginning of retirement and the time of the 2012–13 survey. Again, the biggest increase occurred for cohort 3. People who were (or expect to be) self funded at retirement may actually depend (or expect to depend) on a government pension for the majority of their retirement.

Figure 5 Main source of income


This article has explored the relationship between expectations before retirement, and actual experience of retirement. The focus has been on the timing of retirement, including transitions to retirement, and on main source of income in retirement. Cohort analysis indicated that many of these trends were similar for each cohort over time, but allowed clearer observation of the impact of time/aging on retirement behaviours. These patterns would not have been as clearly observed by analysing differences in the responses of consecutive cohorts at the same age, or different age groups at the same point in time (cross-sectional analysis).

It was found that in general, less people were retired in 2012–13 than expected to be when asked in 2004–05, which indicates that people are retiring later than previously intended. While people are remaining in the labour force longer than expected, the proportion of people working part-time before retirement had increased, and there was a decrease in the proportion of people working part-time hours but preferring more hours. Together these findings pointed to a voluntary transition to retirement through working part-time. It was also found that people were most likely to have government pension as their main source of income at retirement in 2012–13, despite most of the same cohort expecting to retire mainly on superannuation, but that the expectations and actual experience of cohort 2 were more similar than for older cohorts.


1 While this article clearly shows that retirement outcomes are different for each cohort, it is difficult to determine to what extent these differences reflect the effects of age and time period, and which factors are most influential. Even when running a complex regression model, it is not possible to completely separate the effects of these variables.

2 A longitudinal dataset that may provide useful complement to this article is the Household Income and Labour Dynamics in Australia (HILDA) Survey. HILDA is funded by the Department of Social Services (formerly known as FaHCSIA), and run by the Melbourne Institute of Applied Economic and Social Research. For an example of an analysis of retirement intentions using HILDA data, see Cobb-Clarke, D.A. & Stillman, S. (2009) 'The Retirement Expectations of Middle-aged Australians', The Economic Record, 86:269, pp146.

3 English Longitudinal Study of Ageing (Wave 3), 2006, Institute of Fiscal Studies, United Kingdom, http://www.ifs.org.uk/elsa, accessed 4/11/13.

4 The difference is only significant at 0.1 significance level (p-value = 0.094)

5 For example, a recent Australian Human Rights Commission (AHRC) study found that 71% of respondents feel that age discrimination is common in the Australia, 88% of individuals perceive that age discrimination occurs most frequently in the workplace, and 38% of people age 55+ have experienced age discrimination (AHRC, 2013, Fact of fiction? Stereotypes of older Australians, http://www.humanrights.gov.au, accessed 4/11/13).

6 Under employment occurs where a person is employed part-time, but wants and is available to work more hours, or where a person is employed full-time but is temporarily working part-time for involuntary reasons such as being stood down.

7 Recent retirees refers to people who have retired in the last five years. For more information, see Summary of Findings within Retirement and Retirement Intentions 2012–13.