There’s a conversation that happens in every APS HR forum but rarely in public. The classification structure of the service is drifting upward, year by year, in ways that don’t reflect what people are actually doing. Roles that were sensibly EL1 a decade ago are now routinely EL2. Roles that were APS6 are EL1. The implications for budget, capability planning, and the basic credibility of the classification framework are real and underexamined.

This isn’t a new problem. The APS Commission has written about classification creep for years. What’s new is the pace through 2024-25-26 and the structural drivers that have accelerated it.

What the numbers show

The Australian Public Service Commission’s State of the Service Report tracks classification distribution over time. The pattern across the most recent reports is clear. Lower classifications (APS1, APS2, APS3) have continued to shrink as a proportion of the workforce. Middle classifications (APS4, APS5) have been broadly stable. EL classifications and SES have grown.

Some of this reflects genuine work-mix change. The APS has shifted toward more knowledge-intensive, policy-focused, and complex regulatory work over the last two decades. Some of those shifts reasonably push roles up the classification ladder.

A lot of it doesn’t. When you look at the actual work being done at a given level, particularly at the EL1 and EL2 levels, the variance is enormous. There are EL2s in some agencies doing what would be APS6 work in other agencies. There are EL1s doing what would be APS5 work twenty years ago. The classification has drifted while the work has, in many cases, stayed roughly the same.

The structural drivers

A few forces have driven this and none of them are mysterious.

Recruitment pressure. It’s hard to attract people from the private sector or from other agencies into mid-level roles. One way agencies have responded is to push the classification up. An APS6 role that won’t attract candidates gets re-advertised as EL1. The role description shifts slightly but the work doesn’t change much. The salary gets to a level where candidates apply.

Retention pressure. Once people are in the service, advancement is one of the main retention levers. If an agency can’t pay well enough, can’t offer interesting enough work, or can’t compete on flexibility, classification progression becomes the implicit compensation. People get promoted to EL1 or EL2 to keep them, not because the role they’re doing has changed.

Span-of-control inflation. The proliferation of small teams and “manage one person” arrangements has artificially created supervisory roles that justify EL classifications. When a team of three has two team leaders and one operative, something has gone wrong structurally, and yet the pattern is widespread.

Position description drift. When a position is re-advertised after someone has held it for a while, the position description often expands to reflect what that incumbent actually did rather than what the role requires. Subsequent incumbents are recruited against the expanded description even if much of what’s described is incumbent-specific rather than role-specific.

Workplace bargaining incentives. Without naming any specific bargaining outcome, the structure of APS enterprise agreements over multiple cycles has created incentives that flow more strongly through classification progression than through within-classification pay movement.

Why this matters operationally

A few practical consequences.

Budget impact is real. The classification mix has a material effect on agency expenditure even at constant headcount. A 100-person agency with the FY26 classification mix costs measurably more than the same agency with the FY16 mix doing similar work.

Career structure issues. New entrants joining at APS3 or APS4 are looking at career structures where the meaningful “step up” jumps are now to EL2 and SES roles, rather than to EL1 from APS6. The intermediate steps that used to mark a credible career progression have been compressed or eroded.

External recruitment becomes harder. When EL1 in the APS means substantially less than EL1 used to mean, external candidates from comparable private sector roles don’t recognise the proposition. Either they’re offered roles that look junior to them, or they’re offered EL2 roles that are again drifting in the same direction.

Capability planning suffers. When the classification doesn’t reflect the actual work, workforce planning becomes harder. An agency that thinks it has 80 EL2s capable of doing EL2-level work may actually have 40 doing genuine EL2 work and 40 doing something closer to traditional EL1 work. That matters for planning major projects, restructures, or capability shifts.

The Australian National Audit Office and the Productivity Commission have both touched on aspects of this in recent reviews, but no comprehensive analysis has been undertaken since the 2019 Thodey Review.

What can actually be done

The honest answer is: not much in the short term without significant pain. The path that got us here was paved with rational individual decisions and the path back would require accepting some of those decisions need to be reversed, which is politically difficult both inside agencies and across the service.

A few things are possible, though, and several agencies have made progress on them.

Job evaluation discipline. Regular re-evaluation of position descriptions against the work-level standards, with consequences for evaluations that consistently support classification levels not justified by the work.

Role design reform. Looking honestly at whether spans of control, supervisory expectations, and decision authorities actually align with the classifications attached to them.

Career progression alternatives. Building advancement pathways that don’t rely solely on classification progression — specialist tracks, project leadership rotations, deeper expert roles — so the only way to keep good people isn’t to promote them to a role title that doesn’t match the work.

Bargaining outcomes that address structural progression. This is the politically hardest one, and it requires multi-cycle consistency to actually shift the dynamic.

Several agencies have started serious work on at least the first two of these. The early indications are that it’s hard work that bites against everyone’s instincts, but that sustained discipline does produce measurable change over a three-to-five-year horizon.

The honest position

Classification creep isn’t anyone’s fault and is everyone’s responsibility. It accumulated through a thousand reasonable individual decisions, and reversing it requires a thousand inconvenient individual decisions in the other direction. The agencies that get this right will be the ones with credible career structures, manageable budget trajectories, and workforces that can absorb future capability shifts. The agencies that don’t will, eventually, have a classification structure that’s so disconnected from the actual work that something larger has to give. The longer it goes on, the larger that something will be.