Go to top of page

Ch 2: APS 1-6 supply and demand

This chapter continues the examination and analysis of workforce planning risks throughout the APS begun in Chapter 1 Graduate Supply and Demand. Its foci include risks associated with an ageing workforce and changing growth patterns in the workforce. The paper evaluates the APS 1–6 classifications in the same terms, including the extent to which the vacancies created by staff turnover are filled by the internal labour market.

In order to ensure that the Australian Public Service (APS) remains an effective and productive organisation, a professional and engaged APS workforce is essential. While not responsible for developing and focussing the strategic aims of agencies, APS employees are responsible for achieving them. Furthermore, their role in service delivery to citizens and stakeholders is critical in shaping community perceptions of the APS.

Further challenges are presented in Ahead of the Game: Blueprint for the Reform of Australian Government Administration[4]. These are not just restricted to productivity challenges, but include the need for cultural change within the APS. Fostering cultural change and encouraging a “One APS” identity requires a solid understanding of the workforce itself. Knowledge of trends over time is also required to ensure that this change is long-lasting, and not restricted to a single cohort.

The current workforce


Figure 2.1 shows growth rates for the APS 1, APS 2 and APS 3 classifications as a percentage of 1996 numbers based on Australian Public Service Employment Database (APSED) data. All groups show a clear decline from 1996 levels. APS 1 employees were the most affected, with their numbers declining by 70% in three years. Within five years, APS 2 numbers had declined by nearly 50%. Both groups continued to fall in number before rising slightly between 2005 and 2007. By contrast, APS 3 numbers declined significantly before they rose again until 2007. After 2007, all levels again declined.

Figure 2.1: Growth rates for APS 1–3 categories

Figure 2.2 shows the growth rates of the APS 4, APS 5 and APS 6 classifications between 1997 and 2010 as a percentage of 1996 numbers. In contrast to the APS 1 and APS 2 classifications, these groups have shown consistent growth after falls in the late 1990s. All have increased in relation to their 1996 numbers, although the growth rate for the APS 4 classification appears to plateau after 2007.

Figure 2.2: Growth rates for APS 4–6 categories

The conspicuous growth in APS 4 numbers between 1998 and 1999 corresponds to the decrease in APS 3s shown in Figure 2.1. A possible explanation for this is the reduced use of broadband classifications by agencies. Historically, some agencies have reported the classification of employees as bands, including APS 1–3 and APS 4–6. These employees have been recorded in APSED at the highest possible classification until the correct classification has been provided. As such, this change may be due in part to employees being reallocated out of the APS 3 cohort and reported as either APS 1 or APS 2. This would cause an apparent decline in APS 3 numbers. Similarly, other employees may have begun to be reported as APS 4 after previously being reported as APS 5 or APS 6, thus contributing to an apparent growth in this group.

Figure 2.3 shows the numbers of APS 1, 2 and 3 employees as a percentage of the entire APS. While the APS 1 and 2 classifications show a consistent decline in representation, the proportion of the workforce made up by APS 3 employees declines rapidly to approximately 14% between 1996 and 1999. Growth subsequently returns, generally keeping pace with the wider APS until 2006 before dropping again. It is unclear whether this marks the start of a longer, more gradual fall in APS 3 representation, or a temporary decline.

Figure 2.3: APS 1–3 as a percentage of the entire APS

Figure 2.4 shows the numbers of APS 4, 5 and 6 employees as a percentage of the entire APS. While the APS 5 and APS 6 classifications have increased their representation over the last decade, the APS 4 category has shown more dynamic change. As previously discussed, the large and sudden increase in 1999 corresponds to the large decrease in APS 3 numbers at the same time. However, the APS 4 category failed to maintain this apparent rate of growth and has since declined to approximately 20% of the APS. Figure 2.4 also shows that there are considerably more APS 4s and 6s than there are APS 5s. This may have an effect on career paths as large numbers of APS 4s compete for a small number of APS 5 positions. Conversely, the number of APS 5s who are suitable for promotion may be much smaller than the number needed to fill vacant APS 6 positions. This may make the APS 6 cohort more dependent on the external labour market for new employees.

Figure 2.4: APS 4–6 as a percentage of the entire APS

Figure 2.5 shows the increase in the proportion of the workforce which is female since 1996. Despite the different growth rates, classification groups are sufficiently similar in relation to gender that they may be collapsed. All groups have experienced an increase in the proportion of females employed since 1996, which is in line with the larger APS trend. As of 2005, the majority of all employees at these levels were female.

Figure 2.5: Gender breakdown of APS workforce

Figure 2.6 reveals the APS cohorts broken down by age as of July 2010. As shown, the APS 1—2 group has a much higher proportion of Baby Boomers[5] than either of the other classifications. It is unclear whether this represents an older, stable workforce which survived the retrenchments of the late 1990s or another group that has since entered the APS at these lower levels. What is clear, however, is that the APS 1–2 workforce is older than all other APS classifications except EL 2 (60% Baby Boomers) and SES (73.6% Baby Boomers). By contrast, Baby Boomers comprise approximately 38.5% (APS 3–4) and 41.5% (APS 5–6) of the APS workforce respectively.

Figure 2.6: APS workforce broken down by generation

Figures 2.7 to 2.9 show changes in the age profiles of the APS workforce since 1996. Figure 2.7 reveals that the APS 1—2 workforce has aged considerably, with the proportion of employees aged 34 or younger dropping by nearly 20% since 1996. There has also been a corresponding increase in employees aged 55 and over. The APS 3–4 category shows a similar, although less pronounced, trend. By contrast, the APS 5–6 workforce has been relatively stable, with a steady 30% of employees aged 34 or younger. There has, however, been an increase in the proportion of employees aged 55 and over, matched by a corresponding decrease in employees aged between 40 and 44. This suggests a group composed of two parts: a stable, ageing group and a younger cohort that is being constantly refreshed with new employees.

Figure 2.7: Age trends for APS 1–2 employees

Figure 2.8: Age trends for APS 3–4 employees

Figure 2.9: Age trends for APS 5–6 employees

Staff turnover—Rates, reasons, and intentions

Figure 2.10 shows the separation rate as a three year moving average for ongoing APS employees since 1998. A large increase in APS 2 separations occurred in 2005 which is responsible for the apparent rise in APS 1–2 separations. By contrast, APS 3–4 and APS 5–6 separation rates have been more consistent, stabilising at approximately 7% since 2003.

Figure 2.10: Separation rates for APS employees

Separation types

Figures 2.11 to 2.13 show the types of separations by APS employees between 1996 and 2010. For all employees, the large-scale retrenchments of the late 1990s are visible. Since 2001, resignation has been the single largest cause of separation for APS 3–4 and APS 5–6 employees. For APS 1–2 employees, a rise in the numbers of retrenchments and forced moves to a non-APS agency in 2005 interrupts this trend.

Figure 2.11: Separation types for APS 1–2 employees

Figure 2.12: Separation types for APS 3–4 employees

Figure 2.13: Separation types for APS 5–6 employees

All graphs point to an apparent decline in the number of resignations since a peak in 2008 which coincides roughly with the Global Financial Crisis (GFC). A similar, but smaller peak also occurred in retirements in 2008. It is unclear whether this is the beginning of a longer-term trend created by the GFC, or simply a return to the more stable levels seen between 2001 and 2005.

Career intentions

Figure 2.14 shows the career intentions of employees of each classification broken down by generation based on the 2010 State of the Service employee survey. At all levels, Generation X and Y employees are more likely to report intending to leave the APS within the next two years than Baby Boomers are. They are also more likely to report intending to leave their agency.

Figure 2.14: Career intentions for APS employees

Figure 2.15 sets out the proportion of each classification who were eligible to retire at the time of the survey (aged 55 or older) and who intended to do so within the next two years. There are no significant differences between the classifications. In each case, only about one in four employees who are eligible to retire actually intend to do so in the near future.

Figure 2.15: Retirement intentions for APS employees

Recruitment requirements

In order to estimate the APS recruitment needs (Demand), the number of engagements at each level is added to the total number of transfers to that level[6]. The proportion of the Demand composed of internal applicants estimates the dependency of the APS on the external labour market.

Filling the Demand using internal applicants causes transfers or promotions for individuals. This is frequently from the level immediately below the vacancy—its feeder group. For the APS as an organisation, however, this results in the Demand being transferred to the feeder group. Put another way, promoting an APS 5 meets the Demand at the APS 6 level, but increases the Demand at the APS 5 level.

Figures 2.16 and 2.17 contain the proportion of the Demand met by internal applicants for the APS 1–3 and APS 4–6 workforces, respectively.

Figure 2.16: Proportion of APS 1–3 demand filled by internal applicants

Figure 2.17: Proportion of APS 4–6 liability filled by internal applicants

As Figure 2.16 shows, the proportion of APS 1 positions filled by internal applicants has been consistently low since the major workforce changes of the late 1990s. This makes intuitive sense, as this group has no feeder group from which to draw new employees. The dependence of the APS 2 and APS 3 groups on the external labour market has increased since 2000, probably because of the reduction in size of their feeder groups. However, this did lessen somewhat between 2007 and 2009. This was largely due to a peak in the raw number of successful internal applicants rather than a fall in Liability. This suggests that despite the shrinkage of feeder groups, Liability should be able to return to these peak levels without an increased APS dependency on external recruitment.

The dependence of the APS 4 and APS 6 classifications on the external labour market has always been relatively low. Since 2003, 80% or more of the Liability has been filled by internal applicants. This has not been affected by changes in the size of their feeder groups. The APS 5 classification has always been more dependent on the external labour market, although the extent of this has varied. Over time, the number of APS 4 employees relative to APS 5 employees has fluctuated (see Figure 2.4), but again the changing size of the feeder group has not affected recruitment. It seems likely, however, that Liability can return to peak levels without an increasing APS reliance on the external labour market for APS 5 employees.

Workforce modelling

Separation and promotion rates have remained relatively stable for each classification since 2000. However, over the last 10 years, the number of APS 1–2 employees has fallen while the number of APS 3–6 employees has grown. There may be a relationship between these trends as the more routine tasks become automated and a greater number of customer requests are dealt with online, thus reducing the need for traditional, service delivery positions in favour of more skilled staff. While large fluctuations in annual growth rates make forecasting difficult, it is likely that these trends will continue unless other factors change them. These may include reaching a minimum number of APS 1–2 employees, below which it becomes impossible to deliver services effectively.

Given the key role of growth in determining liability, modelling will combine the two negative growth categories (APS 1 and 2) and the positive growth categories (APS 3–6). Two possible future scenarios will be modelled:

Scenario 1: Continuing positive and negative growth for different classifications.

APS 1–2 numbers will continue to decline while APS 3–6 numbers grow as the nature of APS work becomes more technical.

Scenario 2: Negative growth for APS 1–2 ceases after 2013 while positive growth continues for APS 3–6 at a slower rate.

The negative growth in the APS 1–2 classifications will halve its current levels in 2012 and become 0% in 2013. After two years, APS 3–6 positive growth will be reduced by half as changes to the workforce near completion.

Scenario 2 assumes that within three years the APS workforce will have largely eliminated the need for APS 1–3 employees as the move to online service delivery and electronic records management nears completion, thereby automating the tasks usually performed by these employees. Replacing them with more senior technical experts and analysts will have contributed to the growth experienced in the APS 4–6 classifications. As this process is completed, APS 4—6 growth will slow.

Full results of the modelling can be found in Annex A. Summarised results can be found in Figures 2.18 (APS 1–2) and 2.19 (APS 3–6).

Figure 2.18: Results of modelling for APS 1–2 employees

Figure 2.19: Results of modelling for APS 3–6 employees


This work is limited in a number of important ways. As with all modelling, it is based on historical data. Estimates are only accurate to the degree the future is the same as the past. Changes to government policy or the economic situation will threaten the validity of the modelling.

Specifically, this modelling assumes that separation rates will remain stable. As one of the primary sources of Liability in the APS, should rates change this modelling will become invalid. Furthermore, as the oldest workforce category in the APS, the APS 1–2 group are particularly vulnerable to an increase in separation rates as the number of retirements increases.

A key limitation to this report is that it is based on a head count of the APS, not a Full Time Equivalent (FTE) worker approach. The capacity of the APS may not have changed in line with the changing numbers of employees if employees are part-time. If there has been an increase in the use of part-time and job sharing arrangements in the APS, particularly at the APS 1–3 levels, the use of a simple head count approach may over-estimate the capacity of the APS.


Different parts of the APS workforce have grown at different rates since 1996. While all APS classifications were affected by the large-scale retrenchments of the late 1990s, the APS 1 and 2 classifications have continued to decline in both real numbers and as a proportion of the entire APS. By contrast, the APS 4–6 categories have experienced considerable growth since 2000, while the APS 3 classification has stabilised.

There are also demographic differences between the classifications. Most notably, APS 1 and APS 2 employees tend to be older than other classifications. This may have implications for workforce planning, as these groups are more vulnerable to ageing issues, including a higher incidence of chronic health conditions, and their eligibility to retire. Unless APS 1 and APS 2 numbers are reduced, it will be necessary to replace departing employees.

While there has been a decline in resignations since the peak of 2008, there is not yet sufficient data to say whether this is the start of a larger trend caused by the GFC, or a return to historically stable levels.

The reduction in size of the APS 1–3 workforces may have lead to a systemic change in the APS career path, shifting the most frequent point of entry upwards. If this is the case, then the decline in numbers of APS 1–3 positions may not have had a major impact on the career paths of individual public servants. While this paper has not examined this issue, it is also possible that this has been accompanied by an increase in the proportion of new entrants with tertiary qualifications. This may have been part of a longer-term, societal trend towards higher education.

There is a complex relationship between demographic and organisational factors. While APS 3 employees may be demographically similar to APS 4 employees, the growth rates and other organisational factors might be substantially different between the two. When looking at the APS 1-6 workforce, a decision should be made as to whether the primary focus is the demographics of the employees or the changes to the classifications themselves.


4 Advisory Group on Reform of Australian Government Administration, Ahead of the Game: Blueprint for the Reform of Australian Government Administration, Commonwealth of Australia, Canberra, 2010. (Back to content)

5 Baby Boomers are employees aged over 45 in 2010. By contrast, Generation X refers to employees aged between 30 and 44 in 2010, while Generation Y refers to those aged 29 or younger. (Back to content)

6This Demand figure approximates the Supply of new workers. An alternative is to use the increase in the workforce at each level combined with the number of separations to estimate the Demand for new staff. While Supply should approximate Demand there appears to be some discrepancy in APSED with historical data. This is because the demand figure is affected by a number of sources of untracked movement in the APS, including reallocations caused by the increasingly accurate capture of member level. The Supply estimate is not as affected by these issues and is likely to be a more reliable estimate, as the APS only draws on the labour market to fill these requirements. (Back to content)

Last reviewed: 
21 May 2018