Australian Federal Police: Sixteen Years of the Same Question

The AFP, Accountability, and the Persistence of Discretionary Impunity

In 2010, an AFP officer told Parliament his own organisation could not define corruption. In 2026, that same organisation has closed more than half of all misconduct allegations against it without ever investigating them. What, exactly, has changed?

I. A Question Asked In 2009

On 6 December 2010, I published an essay on this platform prompted by a single admission. Paul Jevtovic – then Acting Chief Operating Officer of the Australian Federal Police, later an assistant commissioner – had told a parliamentary inquiry that there was no universally agreed definition of corruption, and that a clearer statutory definition would help the AFP police itself. I treated the admission as what it was: an organisation asking Parliament to tell it what its own members were forbidden from doing, well into its institutional adulthood.

I set that admission alongside two other data points from the same period. The first was Commonwealth Ombudsman Professor John McMillan’s account, given to Four Corners in October 2008, of an office that found the AFP more resistant to oversight than any other Commonwealth agency, and that had been told by senior officers it was wasting police time on trivialities. The second was the case of Mark Standen, the former senior investigator with the NSW Crime Commission – an organisation staffed substantially by AFP alumni – who had been arrested in 2008 and charged with conspiring to import a commercial quantity of a chemical precursor for the drug ice. I also drew on Professor Michael Kennedy’s long record of documenting AFP culture: officers who had stolen and dealt seized drugs, allegations of leaked operational detail, and a complaints architecture in which an AFP officer who found something rotten in his own ranks had nowhere independent to take it, because the Commonwealth Ombudsman was obliged to ask the AFP itself to help investigate the AFP.

My conclusion in 2010 was blunt, and I stand by it: an organisation that cannot define corruption, that treats its own watchdog as an obstacle, and that recycles its personnel through state bodies with equally poor accountability, is not suffering from an isolated cultural problem. It is describing, in its own words, a structure built to resist scrutiny.

II. Sixteen Years On: The Same Admission, Dressed in New Statistics

Nearly sixteen years later, on 25 September 2026, figures tendered to the parliamentary inquiry into racism, hate and violence directed at First Nations people gave that structure a number. Since July 2021, 512 formal complaints against AFP officers generated 1,213 individual allegations of misconduct. Of those, 643 – more than half – were closed under section 40TF of the Australian Federal Police Act, a broad discretionary power that allows the AFP to “finalise” a complaint without ever investigating it. Of the remainder, 283 are still under investigation, 63 have been established, and 215 were investigated and not established. 9News’s reporting on the figures, drawing on the same inquiry material, described the pattern as one critics have called a protection racket – a description the underlying numbers do little to dispel.

This is not a rogue statistic. The Commonwealth Ombudsman’s own office – the same office McMillan once described as institutionally under siege from the AFP – has reviewed the AFP’s use of section 40TF on multiple occasions and found the same pattern each time: complaints closed without the relevant evidence ever being identified or reviewed, and decisions made without the documented reasoning that would let anyone check them later. The Ombudsman’s most recent annual review warns that overuse of the discretion risks denying procedural fairness and deprives the AFP of opportunities to improve. That is a regulator’s careful language for what my 2010 essay called, less carefully, an odious refusal to be examined.

The parallel to Jevtovic’s 2009 testimony is almost too neat. In 2009, the AFP told Parliament it needed a better definition of corruption before it could police it. In 2026, the AFP is using a lawful but loosely bounded discretionary power to avoid the question of misconduct altogether – not because the definition is unclear, but because the power to decide an allegation doesn’t warrant investigation sits, unreviewed, with the organisation accused. The 2009 admission was about definitional ambiguity. The 2026 figures are about something more calculated: an organisation that knows exactly what it is being accused of, and has a lawful mechanism to make the accusation disappear before the evidence is even looked at.

III. The Machinery Of Non-Accountability: Section 40TF

It is worth being precise about what section 40TF actually permits, because precision is where these stories are usually lost. The power exists, in principle, for good reason: not every complaint against a police officer warrants a full investigation, and a complaints system that cannot triage frivolous or vexatious matters would collapse under its own volume. Used as intended, discretionary finalisation is an efficiency mechanism, not a cover-up mechanism. That is the AFP’s own defence, and it is not, on its face, unreasonable.

The difficulty – and the Ombudsman has said as much across several reviews, not once – is that the power has been used far beyond triage. More than half of all allegations over five years is not a filter catching obviously baseless complaints; it is the default disposal method. And the Ombudsman’s finding that complaints were closed without evidence being identified or reviewed, and without documented reasons, means the discretion is being exercised in exactly the opaque way that makes external verification impossible. A power that cannot be checked because its exercise leaves no reviewable trail is not a triage tool. It is a trapdoor.

This is the same structural point Kennedy was making to journalists in 2009 about the AFP’s internal architecture, just relocated from the informal – an organisation too young and too insular to handle its own rot – to the statutory: a formal power, lawfully conferred, that produces the same practical effect his informal description warned about. The mechanism has become more sophisticated. The outcome has not changed at all.

It is also worth noting what section 40TF is not. It is not an emergency power, invoked rarely and under exceptional pressure. It is a routine administrative tool, applied as a matter of course to the majority of complaints an external member of the public bothers to lodge. That distinction matters for how we should read the Ombudsman’s repeated interventions. A regulator does not write successive annual reviews flagging the same concern about a genuinely exceptional power; it writes them about a power that has become the institution’s default posture toward being questioned at all. Read across several review cycles rather than in isolation, the Ombudsman’s language has been remarkably consistent: evidence not identified, reasons not documented, discretion exercised too readily. That consistency is itself a finding. An institution capable of correcting a genuine anomaly in its complaint-handling would have produced, at some point across five years and multiple ombudsman reviews, a visible change in the underlying ratio. It has not.

IV. The First Nations Dimension

The context in which these 2026 figures surfaced matters, and it is not incidental to Blak and Black’s interest in this story. The figures were not volunteered by the AFP; they were extracted through questions on notice to a parliamentary inquiry specifically examining racism, hate and violence directed at Aboriginal and Torres Strait Islander people. Within that inquiry’s narrower brief, the AFP disclosed 25 formal complaints of racist conduct by its officers over the same five years, comprising 47 individual allegations. Thirteen of those were closed under section 40TF without investigation. Eight were investigated and not established. Eighteen remain open. Seven allegations – relating to racist language and behaviour at work, including in group chats – were established, and the AFP says those resulted in disciplinary action: remedial retraining for five officers, and formal show-cause notices under section 28 for two.

Independent Senator Lidia Thorpe, giving evidence to the same inquiry, made two points worth taking separately because they operate on different registers. The first is structural: she has called for an independent national body with genuine independence from police to replace the current oversight architecture, and for the legislation underpinning section 40TF to be reviewed so that complainants have a route that isn’t simply a dismissal by the AFP’s own leadership. That is a policy argument, and a defensible one – it is, in substance, the same argument the Ombudsman’s own reviews have been making in more measured institutional language for years.

Thorpe’s second point is empirical and harder to verify from outside: her view that the true number of racism complaints is significantly higher than the AFP’s own figures suggest, because Aboriginal and Torres Strait Islander people who complain about police conduct risk becoming targets for further attention from the same police. I want to steelman the AFP’s likely response here, because tonal discipline requires it: an organisation cannot be held accountable for complaints that are never made, and a chilling effect on complaints, if real, is a policy problem for the justice system broadly rather than proof of AFP misconduct specifically. That is a fair distinction. But it does not survive contact with the AFP’s own five-year figures, which show that even the complaints that were made – properly lodged, properly numbered – were more likely than not to be closed without the evidence being reviewed. If the front door is this unreliable for the complaints that do arrive, Thorpe’s argument about the chilling effect on the complaints that never do becomes considerably harder to dismiss.

V. Continuities: Standen, Marina and the Culture Jevtovic Couldn’t Define

Readers of this platform’s longer-running thread on ACT Policing and AFP accountability will recognise the shape of this pattern, because I have documented variations of it before, under different names, for over a decade. The 2002 letter written by Angel Marina to the ACT Treasury’s chief executive, and the subsequent 2003 letters to the ANZ Bank, the Commonwealth Attorney-General’s Department, Bird Cameron chartered accountants and the University of Sydney, sit in the same institutional terrain as the cases discussed here: allegations made, an internal process controlled by the institution accused, and outcomes that were never subjected to the kind of external, evidence-tested review that would let an outsider assess whether the internal process was fair. Mark Mullins, the target of those letters and a cousin of Aunty Isobel Coe, experienced the same asymmetry that the 2026 figures now describe in aggregate: an institution investigating itself, on its own terms, with no obligation to show its working.

Mark Standen’s case belongs in this genealogy for a different reason. His prosecution, and his subsequent conviction for conspiring to import a precursor chemical used in the manufacture of methamphetamine, demonstrated that the personnel pipeline between the AFP and state-level law enforcement and crime bodies is not a hypothetical vulnerability – it is a documented one, realised at a senior level, in an organisation that shared personnel, culture, and institutional reflexes with the AFP itself. Standen was not an AFP officer at the time of his arrest, but the concern Kennedy raised in 2009 – that the AFP’s own alumni populate the accountability-poor corners of the broader law enforcement ecosystem – was not resolved by his prosecution. It was, if anything, illustrated by it.

None of this is to suggest that every AFP officer, or even most, share the culture these cases expose. Fairness requires saying so plainly, and the AFP’s own account of the 2026 figures makes the same point: seven of the racism allegations were established and led to real disciplinary consequences, including two show-cause notices under section 28 – a mechanism that can end a career. An organisation that closes complaints wholesale without investigation is not, by that fact alone, an organisation incapable of ever holding anyone to account. The two findings sit together uncomfortably, but they are both true: the AFP has, on the evidence of its own disclosures, disciplined officers for established racist conduct, and it has also used a discretionary power to avoid finding out whether hundreds of other allegations had merit at all. The second fact does not cancel the first. It does, however, make the first much harder to trust as representative, because we have no way of knowing how many of the 643 discretionarily closed allegations would have produced a similar finding had anyone looked.

VI. Why Nothing Changes: The Structural Diagnosis

The question my 2010 essay asked, in its closing lines, was rhetorical: if the AFP itself cannot say what corruption is, where does that leave the average Australian at its mercy? Sixteen years on, the more useful question is not definitional but structural, and it has a less rhetorical answer.

Nothing has changed because the incentive structure that produced the 2009 admission has not changed. The Commonwealth Ombudsman remains the primary external check on AFP complaint-handling, and the Ombudsman remains, as McMillan described in 2008, an office whose findings the AFP can absorb, note, and continue past – because the Ombudsman can recommend, review, and criticise, but cannot compel a different outcome in any individual case, and cannot itself reopen a matter the AFP has discretionarily closed. The National Anti-Corruption Commission, established since my 2010 essay was written, has jurisdiction over Commonwealth corruption generally, but section 40TF closures are not corruption findings; they are administrative non-decisions, and a non-decision is a poor target for a body built to investigate decisions. The gap Jevtovic’s testimony exposed in 2009 – a definitional and jurisdictional space where AFP conduct could be neither clearly named as corrupt nor clearly subjected to independent review – has simply migrated to a new statutory address. It used to live in the absence of a definition. It now lives in the breadth of a discretion.

This is why Thorpe’s proposed remedy – an independent oversight body with genuine independence from police, combined with legislative review of the discretion itself – deserves to be taken seriously rather than filed as a familiar political demand. The Ombudsman’s repeated findings over multiple review cycles show that internal correction has been tried, in the mildest possible form, and has not moved the underlying statistic. More than half of all allegations closed without investigation is not a number that improves through further ombudsman commentary. It improves through removing the AFP’s own discretion to decide, unreviewed, that an allegation against it does not need to be looked at.

VII. What Would Actually Change It

I want to close by resisting the temptation – one I indulged more freely in 2010 – to end on outrage rather than specificity, because specificity is what this subject has always lacked from the institutions responsible for it. Three things would move the 2026 figures in a direction the Ombudsman’s reviews have been gesturing toward for years.

First, section 40TF closures should require the same documented evidentiary basis the Ombudsman has repeatedly found missing: a recorded account of what evidence was reviewed, and why it was judged sufficient to close the matter without further investigation, available to the Ombudsman as a matter of course rather than on request. A discretion that must show its reasoning is a materially different power from one that need not.

Second, the discretion to close a complaint under section 40TF should not rest with the AFP alone. An external party – the Ombudsman, or a body of the kind Thorpe has proposed – should have a genuine veto over closures above some threshold of seriousness, particularly for allegations involving excessive force or discriminatory conduct, the two categories the Ombudsman has specifically flagged as being disproportionately closed under the provision.

Third, and this is the point closest to Blak and Black’s founding purpose: any oversight reform needs a specific, measurable commitment to the concern Thorpe raised about underreporting by Aboriginal and Torres Strait Islander complainants. A structural fix to section 40TF that does nothing to address the fear of retaliation that keeps complaints from being lodged in the first place will produce a cleaner set of numbers for a smaller and still unrepresentative sample.

None of these three changes is radical. They are, in substance, what the Ombudsman has been recommending in careful institutional prose across successive reviews, and what a Senate inquiry witness has now said in plainer language on the record. What they require is not a new insight – the insight was available in 2009, and I published it in 2010 – but the political will to remove a self-administered discretion from the organisation that benefits from it.

Conclusion

I do not think it is cynicism to say that a sixteen-year gap between “we don’t know what corruption is” and “we closed 643 of 1,213 allegations without finding out” represents institutional continuity rather than institutional drift. Drift implies a slow movement away from an original position. What the 2026 figures show is an organisation occupying, with greater statutory sophistication, exactly the position it occupied in 2009: functionally immune from having its own conduct tested, for reasons that are lawful, procedurally defensible in isolation, and cumulatively indistinguishable from the grey area Assistant Commissioner Jevtovic once asked Parliament to define for him.

Parliament gave the AFP the discretion it has spent the years since exercising. It can, if it chooses, take a meaningfully narrower version of that discretion back. Whether the inquiry due to report by the end of this year recommends exactly that will be the clearest available test of whether anything, in sixteen years, has actually changed.

I published the 2010 essay under the belief, naive in retrospect, that naming the pattern clearly enough might be sufficient to disturb it – that an organisation confronted publicly with its own admission of definitional confusion would feel some institutional pressure to resolve it. Sixteen years of subsequent reporting, several Ombudsman reviews, a new anti-corruption commission, and now a parliamentary racism inquiry’s worth of hard figures suggest that naming a pattern and disturbing it are two entirely different acts, and that only one of them has occurred. The essay this platform published in 2010 asked a question. The figures released in 2026 are, in every respect that matters, an answer – just not the one the intervening sixteen years should have produced.

– – –

Bakchos is the founder of Blak and Black, an Australian media and advocacy platform established in 2010. Bakchos writes from the intersecting perspectives of Wiradjuri heritage, Jewish identity, and humanism.

© Bakchos, September 2026

BLAK AND BLACK  |  MEDIA AND ADVOCACY  |  EST. 2010