Go to top of page

Ch 4: SES workforce trends

This chapter considers the workforce planning risks that might arise in relation to the often repeated observation that a sizeable proportion of the Senior Executive Service (SES) workforce is eligible for retirement within the next 10 years. In workforce planning terms, this issue is often portrayed as one of simple workforce supply. It broadens analysis of the subject by examining not only whether the internal supply of SES candidates is sufficient to meet the anticipated SES demand over the next 10 years but also separation intentions of the current SES cohort. It seeks to improve understanding of the APS leadership cadre, the impact of an ageing APS workforce, and of this critical element of APS organisational capability.

A large proportion of the APS workforce will reach notional retirement age (i.e. , 55 years) over the next decade. Employees in the 45 and over age group, who will be eligible for retirement within the next 10 years, account for 43.4% of ongoing employees, up from 30.6% in 1996. The ageing of the cohort at more senior classifications, and with longer lengths of service, is particularly evident: for example, at June 2010, 20.5% of SES and 14.3% of ELs were aged 55 and over, compared with 10.7% and 6.4% in 1996. The percentage of the SES eligible for retirement within the next five years is 49.1%.

Given the challenges outlined in Ahead of the Game: Blueprint for the Reform of Australian Government Administration[14], the continued supply of quality SES candidates will be critical to future APS performance. However, the APS, like other employers, has to function in an increasingly tight labour market. In particular, it is likely that high-quality candidates for executive positions will be in demand in both the public and private sectors. And while the APS regularly refreshes the SES cohort from the external labour market it largely draws on its substantial internal supply to fill SES positions.

This paper examines whether the internal supply of potential SES candidates will be sufficient to meet anticipated SES recruitment requirements over the next 10 years. It uses data from the Australian Public Service Employment Database (APSED) to model SES recruitment patterns and the State of the Service Employee Survey to better understand the retirement intentions of the existing SES cohort. To aid modelling and analysis, the data for SES Bands 1 through 3 have been collapsed into a single SES group and, where appropriate, comparisons with the EL2 cohort have also been included.

SES numbers

The SES represents less than 2% of the APS workforce[15] but has been increasing in absolute numbers and as a proportion of the APS workforce during the past decade. However, this has slowed over the last three years. The growth rate in the EL2 workforce during the same period has been similar but has not shown any signs of decreasing (Figure 4.1). Importantly, the absolute number of EL2s is growing more quickly than the number of SES, thus increasing the internal supply of potential SES employees.


Figure 4.1: Senior staff growth rates[16] (firstt panel) and absolute numbers (second panel).

SES age profile

The SES workforce (like the APS workforce) is ageing; Figure 4.2 sets out the age profile of this workforce. This figure reveals the gradual decline in the proportion of SES employees aged less than 55, a recent increase in those aged between 55 and 59, which has levelled off over the past five years, and a steady increase in those aged over 60.

Figure 4.2: Age profile of the SES workforce

Put another way, those SES employees over notional retiring age (55 years) represent almost 20% of the SES workforce, while those approaching retirement age (50–54 ) make up another 27% of the SES workforce. This represents a significant workforce risk to the APS but the dimensions and character of this are not well understood, and consequently an appropriate response to mitigating the risk is not yet clear.

It should also be noted that, while the issue of SES retirement is often portrayed as one of simple workforce supply, that is, ‘Do we have enough numbers?’, the more correct question to ask is ‘Do we have enough numbers of the required quality?’ The issue of quality is beyond the scope of this paper. However, SES talent management was a significant reform focus of Ahead of the Game and one which is being progressed by the Australian Public Service Commission (APSC).

SES separations

Figure 4.3 reveals that, after the very high rates of SES separations in the late 1990s, which were largely due to the policy of the government of the day, SES separations have stabilised at around 6–7% per annum. [17] However, there has been a steady increase in SES separation rates over the last three years averaging 0.34% per annum. Over the same period, EL2 separations have followed a similar pattern but without the recent increase in separation rates, suggesting that there may be an emerging pattern of separation behaviour specific to the SES.

Figure 4.3. Senior staff separation rates

Reasons for SES separations

The three main reasons for SES employee separations are resignation, age retirement and retrenchment; these are plotted in Figure 4.4 over a 15 year period.

Figure 4.4: Senior staff separation types

Striking patterns emerge from this data; after peaking in the 1990s retrenchments declined while there was a concomitant increase in retirements and resignations. Resignations subsequently stabilised, but there has been a steady and continuing increase in age retirements over the last 10 years.

In the last few years there have been dramatic variations in all three separation types characterised by a peak in retirements and resignations in 2008 followed by a substantial decline in 2009 and some signs of recovery in 2010. Given the timing, this effect is almost certainly due to the GFC. Interestingly, there was an almost identical pattern in retrenchments, which lagged 12 months behind the other types of separation, peaking in 2009.

It would appear that, in response to wider economic uncertainty, SES employees delayed retirement decisions and chose the security of existing APS employment over other employment options. While the actual cause of the concurrent increase in agency heads’ use of the provisions of Section 37 of the Public Service Act 1999 (the Act) is impossible to determine from these data, one could hypothesise that it was, at least in part, in response to the changes in SES retirement and resignation behaviour [18].

While the ‘Incentive to retire’ provision of the Act allows agency heads to manage organisational changes that might result from changing priorities, it also provides them with a management tool which enables them to refresh the SES cohort in periods of low turnover. Figure 4.4 shows substantial use of this during the APS downsizing of the late 1990s, and the pattern of retrenchments has tended to follow the rise and fall of SES resignations.

It could be that agencies use Incentive to retire to actively manage their SES workforce in response to changes in employee-initiated separations. Were this the case, the type of SES separation would require ongoing monitoring in order to develop a better understanding both of separation behaviour within the SES and agency approaches to managing the SES cohort.

While these dramatic changes might raise a number of questions about the short-term patterns in the SES workforce, some clear long-term patterns of separations emerge when these short-term variations are compensated for. Figure 4.5 shows the type of separation data as a three year moving average.

Figure 4.5: SES separation types (three year moving average)

From the data presented in Figure 4.5 it is clear that, over the past three years, resignations have plateaued (and may even be decreasing) while age retirements have shown a sustained increase since the early part of the last decade that appears to have been largely unaffected by the GFC. Resignations have also shown a sustained increase albeit only over the past three years. It might be hypothesised that the GFC had a clear effect on the resignation rate although the effect on age retirement behaviour is less clear.

Retirement behaviour: Impact of the CSS

A potential complicating factor in the retirement behaviour of the SES workforce is the impact of membership of the CSS which closed in 1990. The CSS provides a substantial incentive for most scheme members to resign prior to reaching the age of 55, preserve their superannuation, and access their pension at 55 (a process generally known as ‘54/11’). The alternative for these CSS members is to continue working for at least another five years and retire at 60 or more to access the same rate of preserved pension.

While other reviews of the APS workforce noted that employees in the newer Public Sector Superannuation Scheme (PSS) also demonstrated a strong tendency to retire in their mid-50s[19], it remains to be seen if this is a ‘hangover’ retirement expectation for those employed at that time or a more enduring feature of the workforce participation behaviour of older SES employees.

It is possible that this expectation will naturally and gradually diminish as the workforce adjusts to the new superannuation conditions. That is, the expectation of retirement at around 54 will diminish as a part of APS culture with the removal of the CSS ‘trigger’. However, it is also possible that the economic uncertainty generated by the GFC sharpened the focus of the SES cohort on their retirement provisions and altered separation behaviour.

Currently[20], only 35% of the SES are members of the CSS (31% of Band 1, 44% of Band 2 and 54% of Band 3), and only 7% of non-SES employees are members of the CSS. As a result, membership of the CSS may have an impact on more senior members of the SES. However, although there may be some form of enduring expectation of early retirement associated with the CSS, the majority of the SES do not have the same financial imperatives to act on this expectation. Given the low levels of CSS membership among non-SES employees, it seems likely that any effect of this will in fact diminish over time.

Retirement behaviour: Impact of the GFC

In relation to altered patterns of separation, one ‘wildcard’ hypothesis is that the GFC may turn out to have been a major shock that permanently and fundamentally alters the separation (or retirement) behaviour of APS employees. In essence,, an external event that moves expectations about age retirement away from the mid-fifties toward expectations of a longer working life. It may be that the early evidence for this will not be found among the SES cohort where retirement expectations are more likely to be reasonably well-established. Initially, the evidence is more likely to appear in emerging retirement trends across the broader APS that will subsequently be reflected in future changes in SES separation behaviour.

Although the GFC hypothesis seems valid, other factors such as the increase in longevity in the population coupled with the growing realisation of the health benefits of working longer may lead to an alternative (and more positive) hypothesis. It is that, rather that ceasing to work at ‘retiring age’ in a traditional sense they are choosing to participate in work as they age and more than ever before are accessing the full range of employment arrangements to achieve this. In doing so, the patterns of work in the APS are becoming increasingly diverse and this trend is likely to accelerate.

Early support for this hypothesis might be found in changing patterns of part-time and non-ongoing APS employment as well as the increasing engagement of mature age workers. For example, of the 13,725 non-ongoing employees at June 2010, 2,207 or 16.1% had previously worked in the APS as ongoing employees; previous ongoing experience was particularly high among older non-ongoing employees: 49.0% of those in the 55–59 age group, and 49.8% of those in the 60 and over age group had previously worked as ongoing APS employees[21].

Additionally, and in a broader total APS workforce sense[22], some older APS employees are also making the transition to the private sector and continuing to contribute to the APS from beyond the bounds of APS direct employment. This trend cannot be assessed from APS data sets, but it is likely that former SES employees would be well-positioned in terms of skills and experience to make this transition.

As agencies continue to improve the range of employment options they offer mature age employees, the opportunities for the individual to tailor employment to suit their age and lifestyle expectations will improve, as will social acceptance of an employee’s decision to redesign their working arrangements in order to meet late-career needs. Growth in this trend would be the final signal that a relatively long-held APS cultural expectation of full-time retirement in the mid-fifties is no longer valid.

Meeting demand

Table 1 shows annual APS demand for SES employees over the past five years broken down into the number required to meet the needs of a growing SES workforce and the number required to replace departing SES employees. The table also reveals how much of this demand was met from the internal labour market (i.e. , by promotion of EL2 employees to the SES).

Table 4.1. SES supply and demand
SES 2006[23] 2007 2008 2009 2010
Demand 358 375 339 294 213
- Growth 222 244 147 115 41
- Separations 136 131 192 179 172
Internal supply 70.25% 63.25% 74.84% 62.03% 66.95%

The key feature of this data is that the reduction in demand for SES employees is almost totally driven by the slowing of the growth of the SES workforce[24]; this can be seen quite clearly in Figure 4.6.

Figure 4.6: SES staff demand

The other key observation from this is that the source of replacement for the SES is largely independent of the size of the demand. Although only based on limited data, there appears to be no trend in the rate of replacements sourced from the internal labour market which averages around 67%.

Leaving the APS

Analysis of the SES responses to the annual APS Employee Survey provides some insights into SES separation intentions. This analysis shows that there is no statistically significant difference between SES and non-SES employees in relation to their intention to leave their agency within the next two years.

Not surprisingly, given the older age profile, SES employees are significantly more likely to retire than non-SES. However, those intending to retire within the next two years represent less than 10% of the SES cohort. Thirty-five percent of the SES responding to the survey do not intend to leave within the next two years and a further 25% are unsure. Interestingly, almost as many (22.4%) intend leaving to pursue a job in another APS agency as intend to retire.

In the case of SES of retiring age (i.e. , those aged 55 and over) 40% intend to retire within the next two years, but more than one-quarter (29%) are unsure of their intentions and one-fifth (20.3%) indicate that they have no intention of leaving their agency.

Overall, the SES cohort is no more likely to want to leave the APS than other employees, a fact that is reflected in the similarity of their separation rates. They are more likely to leave to retire but this still represents a minority of the SES workforce.

This analysis of intentions to leave demonstrates the complexity of separation behaviour in an older workforce and the risks associated with making assumptions based on demographic profiles alone. The rather bald statement in the opening paragraph of this paper that the ‘percentage of the SES eligible for retirement within the next five years is 49.1%’ provides little insight into the underlying behaviour of the individuals concerned. A more complete picture of the trend only becomes available by monitoring the relationship between separation intentions, actual separations and the attitudes and motivations of the cohort toward work. In terms of the SES cohort, more research is required to match and analyse these three data sources before a trend can be reliably identified.

A final consideration is that this analysis has used 55 years as the indicator for retirement age as it is currently the superannuation preservation age; however, government changes will seethis increase to 60 years progressively over the next decade or so which is likely to have an influence on separation behaviour.

SES workforce modelling

The caveats of the preceding paragraph notwithstanding, an SES workforce model is presented using the following assumptions:

  • The SES has grown substantially over the past 15 years, and the annual growth rate has been quite variable; a conservative growth figure based on the 15 year average will be used to model potential SES workforce growth[25]. This is shown as Demand Model 1 in Table 2.
  • However, a cap on SES numbers was introduced in June 2010 so an alternative model will be presented which assumes no growth in the SES workforce. This is shown as Demand Model 2 in Table 2.
  • Separation rates for SES fell to just under 6% in 2007 but have been consistently increasing over the past three years at a rate of 0.34% per annum; this figure will be used for modelling increases in SES separations since, as argued earlier, this might, in fact, be an enduring change.
  • Internal supply to meet the SES demand has been and will continue to be stable at an average of 67%.

Table 2 shows the SES workforce model separation rates and demand for the next 10 years as well as the number of replacements required from the internal labour market (the internal supply).

Table 4.2: SES demand
Year 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Separation rate 7.27% 7.61% 7.95% 8.29% 8.63% 8.97% 9.31% 9.65% 9.99% 10.33%
Demand model 1 238 269 279 290 301 312 323 334 345 356
- Internal supply 113 133 139 145 151 157 163 169 175 181
Demand model 2 169 199 208 216 225 234 243 252 261 270
- Internal supply 159 180 187 194 202 209 216 223 231 238

The data modelled in Table 2 show the potential SES demand rising over the next 10 years; however, this doesn’t reach the maximum liability experienced as recently as 2007.

Limitations to this work

There are a number of limitations to this modelling, for example, the estimation of the ongoing separation rate as a simple linear extrapolation. It is clear that SES separation behaviour is complex, and further targeted data collection and analysis is required in order to fully understand the main patterns of behaviour.

The other limitation of this modelling is that there have been some major changes in the last three years:

  • SES growth rates have decreased substantially; and
  • SES separation behaviour appears to have changed dramatically.

Three years is too short a time period to use in identifying any reliable trend, and while the GFC is a likely precipitating event, more data will be needed to further explore the hypotheses proposed in the paper concerning the impact of the GFC on separation behaviour within the SES.

Finally, changes to the external labour market are not considered in this analysis; data is not available about the number of external applicants for SES positions but it is not unreasonable to expect that changes in the external labour market will affect the potential supply of SES candidates. In a similar vein, only a limited analysis of the internal labour market has been considered; for example, further work could be done on the propensity of EL 2s to apply for SES roles as this would help refine the analyses.


The SES has been increasing both in absolute size and as a proportion of the APS workforce over the past 10 years; however, this growth rate has slowed substantially in the last three. Within the SES workforce, one of the clear trends is that the SES is ageing and, in particular, the number and proportion of SES aged over 60 who are remaining in the workplace is increasing.

SES separation rates have varied over the last 15 years, but appear to have stabilised in the past six years at around 7% per annum; this is consistent with trends across the rest of the APS workforce, and it may represent a ‘true’ or baseline figure for SES separations. However, over the past three years there has been a steady, if slight, increase in separations that will require ongoing monitoring.

The types of SES separations have changed noticeably in the last three years; resignations have plateaued, while age retirements and retrenchments have increased markedly. The increase in age retirements is not surprising given the ageing of the SES but the increase in retrenchments could, in part, be a response to the reduction in resignations.

It appears that this change in SES separations may have been driven by the GFC, the greatest impact of which appears to have been to reduce resignations. However, its impact on age retirements is less clear. There are some indications that the GFC, among other factors, might have fundamentally changed the way individuals view work as they age and that indicators of those changes are emerging in the form of changing patterns of SES separation behaviour.

Replacement of SES employees has been mainly by promotion from the internal APS labour market; this has consistently accounted for about 67% of new SES employees and varies little irrespective of changes in demand for SES. One of the salient features of SES demand is the role played by retrenchments (41% of separations in 2010); if meeting SES demand were ever to be a challenge, a lever for addressing this could be adjustment of the rate of retrenchments.

In terms of meeting the SES liability, the EL 2 cohort is the main feeder group and its numbers are increasing faster than those of the SES. Although the EL2 cohort’s separation rate has stabilised during the past three years, it has grown at a rate of 3.24% per annum over the last five years, with the result that the pool of potential candidates for SES is growing at a faster rate than the SES separation rate.

SES separation intentions indicate that SES employees are no more likely to leave the APS than any other segment of the workforce; in addition, their mobility intentions are similar to the rest of the APS workforce. However, two factors unique to the SES influence this:

  • SES employees are more likely to age retire than other segments; and
  • a substantial proportion of SES separations are management-initiated and the rate of these has tended to vary widely over the past decade.

Within these caveats, SES workforce modelling shows that, under two different growth assumptions, the demand for SES employees over the next 10 years will not reach exceed levels that the APS has successfully managed since 2007.


[14] 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.

[15] Data based on ongoing employee numbers taken from the Australian Public Service Statistical Bulletin 2009–10.

[16] Calculated using workforce size in 2000 as the base rate of 100%;so the EL and SES workforces have both increased by about 70% since 2000; for comparison purposes, the entire APS workforce has increased by 46% since 2000.

[17] SES separations at 6–7% per annum are based on a three-year moving average.

[18] Public Service Act 1999 – Section 37, Incentive to retire:

(1) An Agency Head may give a notice in writing to an SES employee in the Agency, stating that the employee will become entitled to a payment of a specified amount if the employee retires within a period specified in the notice.

(2) If the employee retires within the specified period, by notice in writing to the Agency Head: (a) the employee is entitled to be paid the specified amount; and (b) the employee is taken, for all purposes, to have been compulsorily retired from the APS.


19 Management Advisory Committee 2003, Organisational Renewal, Commonwealth of Australia, Canberra, pp. 72–94. (Back to content)

20 Data provided by the Australian Public Service Commission’s Workplace Relations Group, July 2011. (Back to content)

21 Source: APSED. (Back to content)

22The total APS workforce consists of all those who work for the APS in day-to-day delivery and support activities. Ongoing and non-ongoing employees, industry contractors and suppliers are all components of the APS workforce who must be regarded as part of the APS’s total human capital capability. (Back to content)

23 As at 30 Jun 2006 and so on for the rest of the Table. (Back to content)

24 These correlate very highly (r = 0.96). (Back to content)

Last reviewed: 
21 May 2018