Go to top of page

8. Assessment of roles against SES work level standards

SES work level standards

The APSC worked with Mercer (Australia) to develop a set of work level standards that describe the work level requirements of SES roles at each of the three SES classification levels. The development of the standards was overseen by a Reference Group chaired by Mr Beale and comprising the Deputy Public Service Commissioner and senior representatives of a number of APS agencies.

The work level standards reflect the leadership capabilities set out in the SELCF and the ILS. They also seek to distinguish SES roles at each level in terms of span of control, the degree of difficulty of stakeholder management responsibilities, the complexity and ambiguity of the job context and environment, the impact of judgements made and the independence with which roles operate.

A key initial premise agreed for the development of the work level standards was that they would be formative rather than prescriptive. This means that the work level standards do not attempt to describe in detail all the possible components of SES roles at their different levels across the APS, a complex exercise that may never have captured the full diversity of SES roles. Rather they focus on the core components that would need to be taken into account in any SES classification decision. This model will require agencies to exercise judgement in applying the work level standards to individual jobs but also give greater flexibility in applying them across the range of APS roles, including atypical ones.

The Reference Group was used to ensure the descriptors in the work level standards reflect the members’ understanding of the requirements of SES roles and the key distinctions between the three levels and between the SES Band 1 and Executive Level 2 classifications.

The draft version of the work level standards developed with the Reference Group was considered suitable for use as a benchmark for an audit of a sample of SES roles.

The audit approach

The aim of the audit was to gather comparative data on SES roles across the APS to test:

  • whether the requirements of SES roles are consistent across the APS and continue to reflect expectations of appropriate work level standards for the SES and
  • the suitability and ease of application of the draft work level standards across the diversity of SES roles.

The 238 SES roles included in the audit were identified through a random selection process.

The first step in selecting SES roles for the audit was to identify agencies that had shown consistent recent periods of significant SES growth. The APSC’s APSED identified agencies that had experienced at least 30 per cent growth in their SES in the 10 years between July 1999 and June 2009. Very small agencies with less than five SES were eliminated from the list and several large agencies, whose growth could not be determined accurately because of machinery of government changes over the time frame, were added. Finally, a couple of agencies that had experienced little or no SES growth over the period were added for comparative purposes.

This resulted in 30 agencies being selected for the audit. The individual SES roles in these agencies to be assessed were then identified using a random selection from the APSED data base that were representative of the overall size and structure of each agency. The sample was able to cover a wide range of SES functions and responsibilities, although a few small adjustments to the sample were necessary, particularly where roles and functions were found to have changed. Six EL2 roles were also assessed.

The audit was conducted by combined APS/Mercer teams and involved a review of position descriptions (or similar) for the roles, information about reporting relationships and financial and staff management responsibilities and relevant business plan documentation supplied by agencies. It also involved a 30 minute interview during which incumbents were asked a standard set of questions about their role.34 A majority of these interviews were face to face but a limited number of phone interviews were also conducted, mainly when the SES roles were located interstate or overseas. A copy of the audit questionnaire is at Attachment E.

Of the 238 SES roles assessed 43 (30.7 per cent of the total number of SES at that level) were SES Band 3, 51 (8.9 per cent of the total number of SES at that level) were SES Band 2 and 144 (7.17 per cent of the total number of SES at that level) were SES Band 1. The small overall number of SES Band 3 roles in the APS meant that it was necessary to sample a much larger percentage to ensure statistical reliability. An important caveat to the interpretation of results of the audit and its application to the APS as a whole is that the selection of agencies on the basis of those that had experienced significant growth could bias the audit findings – ie those agencies which have had the highest level of change might reasonably be thought to be at greater risk of classification errors than those which have remained relatively stable.

The draft SES work level standards were the primary assessment tool, with the points-based Mercer CED Job Evaluation System, which measures the relative size of roles by measuring the major components of work value, providing a secondary tool of delineation. Overall Mercer found that both methodologies provided consistent classification outcomes. The Job Evaluation System points-based work value ranges corresponding to the three SES levels as described by the work level standards were identified by Mercer as follows:

  • SES Band 1: 685–949
  • SES Band 2: 950–1499
  • SES Band 3: 1500–2179.

The Mercer CED Job Evaluation System is “…designed to measure the relative size of positions. It measures the major components of job worth to achieve this. This well established method examines the complexity of job demands of individual positions in a way that allows a systematic and analytical comparison of positions”.35

Audit findings on the classification of SES roles

Overall the audit found that 89 per cent of the 238 SES roles assessed are correctly classified.

Of the remaining 11 per cent, the audit found that:

  • 8 per cent (20 roles) were over classified and
  • 3 per cent (7 roles) were under classified (or 4 per cent if the three SES Band 3 roles above the notional upper limit of the SES Band 3 range are considered under classified).

The findings in terms of the numbers of roles falling within or outside of the work value points range for their classification are presented below.

Chart 6.1 below provides an indication of the distribution of the overall population of SES roles by extrapolating the distribution of the sample to the population. The work value range for each SES band has been split into quartiles (with an additional section for roles classified above the SES Band 3 range). Again, a significant assumption is that the sample identified for the audit is representative of the APS SES as a whole.

Charts 6.2 to 6.4 identify SES roles that fall within 10% of the point score at the top (“upper”) and bottom (“lower”) of each range.36

It is important to note, in the discussion of classification standards for all three SES bands, that placement of roles at the upper or lower quartiles of each work value range reflects the wide distribution of SES roles across that range and does not necessarily imply that roles have been poorly designed.

SES Band 1

Eight SES Band 1 positions (6 per cent) did not achieve the minimum work value score (685 points) for the SES Band 1 range. Mercer identified the key factors for over classified roles were the breadth and depth of expertise required, and the degree of complexity and novelty, which impact the level of judgement required of the SES Band 1 cohort.

Three per cent of SES Band 1 roles exceeded the maximum score for the SES Band 1 range.

Twenty four per cent of SES Band 1 roles fell within 10 per cent of the bottom of the work value range. According to Mercer, this reflects a strong prevalence of principal specialist positions demanding high expertise requirements in a narrow, specialised field. The outcome may reflect what appears to be an increase in the number of specialist policy advice positions at this level, particularly in the central agencies.

Table 6.1 – SES roles within, below or above range
Classification Classified within range Under classified Over classified Total
SES Band 3 34 (79 per cent) 0 9 (21 per cent) 43
SES Band 2 45 (88 per cent) 3 (6 per cent) 3 (6 per cent) 51
SES Band 1 132 (91 per cent) 4 (3 per cent) 8 (6 per cent) 144
EL 2 6 (100 per cent) 0 0 6

Chart 6.1: Summary of SES role assessment outcomes

A relatively high level (14 per cent) of SES Band 1 roles fell within 10 per cent of the top of the work value points range. These roles exhibited a higher degree of job environment complexity, and a requirement for long-term strategic development and forward planning relative to the core SES Band 1 cohort.

The distribution of SES Band 1 roles across the relevant points range suggests that the distribution of these roles is skewed towards the bottom of the points range, but that they fall disproportionately into the second quartile.

SES Band 2

Mercer observed that the SES Band 2 group reflects the most varied cohort in terms of focus, ranging from significant program management and delivery, high level/high impact policy advice and principal experts.

Three roles (6 per cent) did not achieve the minimum score for the SES Band 2 range and fell at the upper end of the SES Band 1 work value range. Key factors for over classified roles were the degree of job environment complexity and the requirement for long-term strategic development and forward planning relative to the work level standards that is expected of the SES Band 2 cohort.

Three roles (6 per cent) exceeded the maximum score for the SES Band 2 range. These roles had very significant breadth and scale of operations, impacting both agency and community, and a predominant focus on long-term strategic direction.

Twenty three percent of SES Band 2 roles fell within 10 per cent of the bottom of the work value range. An examination of work value trends and patterns indicates that these roles are predominantly program or policy focused and that while they have an appropriate and even high level of program, policy, strategic or operational responsibility and complexity, Mercer observed that it was across a relatively narrower set of issues. It was also noted that the level of expertise and independence in critical decision making processes for programs, policies and operating circumstances was reduced in these roles relative to core SES Band 2 roles. A further influence is the role’s place in the organisational structure and the governance framework that sits alongside the role’s point of focus. The more roles involved in the delivery/design of a particular outcome, the more diluted the work value for each role will become.

As with the SES Band 3s (below), the distribution of SES Band 2 roles across the relevant points range suggests that the distribution of these roles is skewed towards the bottom of the points range.

Chart 6.2: Distribution of SES Band 1 roles

Chart 6.3: Distribution of Band 2 roles

Twenty-one per cent of SES Band 3 roles (9 roles) did not achieve the minimum score for the SES Band 3 range. Key factors for over classified roles were the breadth of focus and complexity of stakeholder management, budget size and impact on agency performance expected of SES Band 3 cohort.

Three SES Band 3 roles exceeded the maximum score for the SES Band 3 range demonstrating work value possibly equivalent to some departmental secretary roles.

Discussion

Because the audit sought to identify broad trends across the APS it focused on the breadth of SES roles assessed rather than the depth of assessments of individual roles, as would be required if strong conclusions were sought for individual SES roles. For example the assessments did not involve an interview with the roles the sampled roles report to or direct reports. The findings are therefore indicative of the broad distribution of work value associated with SES roles but cannot support strong conclusions about individual roles.

The distribution of roles across the work value points ranges are likely to reflect the differences in both the breadth and nature of work within any classification band, differences that will change over time as the complexity and size of particular programs and projects fluctuate.

The finding of a significant number of apparently misclassified roles in the audit is at least primafacie support for the view that that some SES roles are designed with insufficient regard to job design principles or reference to appropriate work level standards. If the assessment process effectively captures classification ‘errors’ and is representative of the total SES population, it would suggest that around 115 SES Band 1 are over classified and 57 under classified; around 33 SES Band 2 are over classified and 33 under classified; and around 26 SES Band 3 are over classified.

If this assessment is accurate the net annual cost to the Commonwealth on the basis of median total remuneration package for each SES classification level would be around $6 million, or about 1.05% of the estimated total annual SES remuneration costs – based on median total remuneration package at each level multiplied by total number of SES. This assumes that both over classified and under classified roles receive the median remuneration package for their level. Also, as noted above, an important caveat to the interpretation of results of the audit and its application to the APS as a whole is that the selection of agencies on the basis of those that had experienced significant growth could bias the audit findings – ie those agencies which have had the highest level of change might reasonably bethought to be at greater risk of classification errors than those which have remained relatively stable.

“Classification creep” – a change in the classification profile over time that is not attributable to changes in the nature of the work being performed – is more likely to occur when there is, among other things, no standardised approach in terms of methodology and where internal (agency specific) benchmarks are a major determinant.

Mercer has suggested that the draft work level standards could form the basis of a more standardised approach to the classification process. However, SES Band 3 roles are usually much broader based, in terms of functions and accountabilities, than those at SES Band 1 and 2 classification levels and the work level standards division of SES roles into the four categories of Delivery, Public Policy, Regulatory and Professional/Specialist is of less relevance at the SES Band 3 level. The Reference Group noted that on some occasions SES Band 3 level jobs are assigned to roles not because of the breadth of the role but the political,financial or community risk associated with them. Sometimes these roles are responsible for complex inter-governmental or public private negotiations and policy development, for major innovation and risk, and for the shepherding through of key national initiatives or for the review of critical areas of policy.

Chart 6.4: Distribution of Band 3 roles

Factors such as these may not have been adequately comprehended by the draft work level standards or Mercer’s points factor system. The work level standards were amended subsequent to the audit to provide greater emphasis on these issues at Band 3.

Over the past 20 years there has been a significant trend for SES Band 3 positions to form a line management role coordinating groups of divisions and functions, rather than the ‘off-line’ strategic and departmental secretary’s ‘alter ego’ role they predominantly played before 1987. This might reflect both the consolidation of departments at that time (which has been more or less maintained) and the growing demands on Secretaries’ time to support Ministers, appear before the Parliament and other inquiries, deal with key stakeholders and participate in international and COAG negotiations.

In its report, Mercer canvassed the issue of a revised four level classification structure, based in part on the assessed high levels of responsibility at the top of SES Band 3. However, this is not sufficient justification for a significant revision of the current classification structure. Consistent with the formative approach agreed by stakeholders at the outset of the development process the work level standards do not and are not designed to reflect fine work value distinctions between roles. Secondary delineation tools such as a work value points factor or a more detailed but not points based system can enhance the identification of these distinctions but agencies would still need to exercise judgments based on specific job and agency knowledge and context in making accurate decisions about the classification of SES roles. This is likely to be particularly the case where roles have elements that appear to equally apply across two levels or where the SES role being assessed is unique, with few comparative features.

Individual agency reports were provided to the 30 agencies involved in the audit of SES roles. Comments received from agency heads have been incorporated into the final draft work level standards at Attachment D.

Factors such as these may not have been adequately comprehended by the draft work level standards or Mercer’s points factor system. The work level standards were amended subsequent to the audit to provide greater emphasis on these issues at Band 3.

Over the past 20 years there has been a significant trend for SES Band 3 positions to form a line management role coordinating groups of divisions and functions, rather than the ‘off-line’ strategic and departmental secretary’s ‘alter ego’ role they predominantly played before 1987. This might reflect both the consolidation of departments at that time (which has been more or less maintained) and the growing demands on Secretaries’ time to support Ministers, appear before the Parliament and other inquiries, deal with key stakeholders and participate in international and COAG negotiations.

In its report, Mercer canvassed the issue of a revised four level classification structure, based in part on the assessed high levels of responsibility at the top of SES Band 3. However, this is not sufficient justification for a significant revision of the current classification structure. Consistent with the formative approach agreed by stakeholders at the outset of the development process the work level standards do not and are not designed to reflect fine work value distinctions between roles. Secondary delineation tools such as a work value points factor or a more detailed but not points based system can enhance the identification of these distinctions but agencies would still need to exercise judgments based on specific job and agency knowledge and context in making accurate decisions about the classification of SES roles. This is likely to be particularly the case where roles have elements that appear to equally apply across two levels or where the SES role being assessed is unique, with few comparative features.

Individual agency reports were provided to the 30 agencies involved in the audit of SES roles. Comments received from agency heads have been incorporated into the final draft work level standards at Attachment D.

Finding 6.1: The work level standards produced by the Review are robust in describing the three SES classification levels, however, for the work level standards to be given efficient and consistent effect, by departments will require development of supporting application guidelines and training.

Finding 6.2: In general the audit suggested that classification decisions are being soundly made and apparent ‘error’ rates fall within normal bounds. However, there is some suggestion that SES Band 3 roles might have a higher rate of ‘over classification’ than other SES classification levels.

Finding 6.3: On the assumption that the audit results can be applied to the total APS population – and there are important caveats in relation to the choice and size of the sample – and if classification ‘errors’ could be corrected the estimated net savings to the Commonwealth are around $6 million pa.

Recommendation 4: APS-wide SES work level standards based on those at Attachment D should be issued by the Commissioner and the Public Service Classification Rules 2000 should be amended to make agency use of the SES work level standards mandatory in classifying SES positions.

Recommendation 5: A simple methodology for evaluating SES roles against the work level standards and a workbook to guide agencies in its use should be issued by the Commissioner – with training of agencies in the use of the methodology to be provided by the APS Commission.

Recommendation 6: Agency heads should be required to evaluate all SES roles against the proposed methodology and the work level standards:

  • for established roles, at a minimum, as they become vacant; and
  • for all new SES roles – including those sought as part of a new policy proposal – evaluations to be based on the standards but include an assessment of the impact of new roles on existing roles and structures.

Recommendation 7: Agency heads certify before a vacant or new role is filled that the classification for the role is based on an evaluation against the work level standards.

Recommendation 8: A risk based audit program of the classification of SES roles be undertaken by the APS Commission.


34 For a small number of roles where information was not supplied by the agency, the assessment was based on the interviews with incumbents

35 Mercer Australia, Audit Report, SES Review, 22 November 2010, (2010:29)

36 For example, the “lower” range for Band 1 roles ranges from 685 points to 753.5 points