The NDIS Bill is argued on fraud. Here is what the police actually told the Senate

This masthead has already set out where the peak bodies stand on this Bill, from “must not pass” to “assume it passes and negotiate the detail”. That was written from submissions and public statements. The hearings are a different kind of source: witnesses are questioned, and the answers are on the record whether or not they are comfortable. Both transcripts are now published, and this is what is in them.

What the police described

Acting Assistant Commissioner Donna Parsons of the AFP’s Crime Command opened by saying the AFP recognises that fraud against the NDIS undermines public confidence, threatens Commonwealth revenue and, most importantly, harms some of Australia’s most vulnerable people.

Adam Meyer, Deputy Chief Executive Officer of the ACIC’s Intelligence Group, described the range: charging for services never provided, overcharging for services that were, and continuing to charge against plans that have expired. The people doing it run from dedicated fraud networks through to bigger, broader organised crime networks who approach targeting of the NDIS as part of a bigger, broader business model, and the money obtained is something the groups then often recycle into a range of different types of other criminal activity.

The part worth reading twice is what he said about participants. Some are knowing participants in kickback arrangements. Others are not, and are coerced:

We know, in some instances, some of the fraudulent providers and actors are threatening violence and a range of other adverse activities against the participants if they don’t comply with what they are suggesting as part of that fraudulent activity.

None of that should be minimised, and this piece does not. It is also the reason the next section matters rather than the reason it does not.

The question the AFP could not answer

Senator Jordon Steele-John asked what data the AFP holds on fraud perpetrated by large providers compared to small providers. Superintendent Timothy Underhill first deferred to the NDIA. Asked again, for the AFP’s own view as an agency deeply integrated into these investigations, he said the AFP collects data on the volumes of fraud it investigates but there’s no system to withdraw data based on the size of a provider, and would need some sort of system for determining what actually constitutes… a large provider versus a small provider.

Pressed a third time, and narrowed to only the tier-one matters the AFP runs itself, the answer was flatter:

I just don’t have that data. In our cases, we don’t have a system where we put in whether they are large or small. That information would be contained within an individual case, but there’s no facility to spit out a statistic on it. There’s no box that’s ticked, ‘This is large or small,’ to enable us to pull out data that would say, ‘This many were large and this many were small.’

The senator’s reply put the objection plainly: the ability to identify patterns of criminal behaviour is a central tenet of law enforcement and investigative process. The matter was taken on notice, with answers due to the committee by close of business on 6 August, eight days before it must report.

Who actually runs these investigations

A second thing emerged that is easy to miss. The AFP works with the NDIA under a service-level agreement with three tiers. Tier one is an AFP-run investigation, reserved for the very highest end of serious offending. Tier three is the NDIA running its own investigation and prosecution, calling on the AFP for search warrants. Tier two sits between them, and on Superintendent Underhill’s evidence it is the vast majority of the matters that we have with the NDIA: the agency runs the investigation, the AFP provides mentoring, guidance and search-warrant capability, and the agency carries the prosecution with the Commonwealth Director of Public Prosecutions.

So when fraud figures are cited in support of this Bill, the body that holds most of the underlying case detail is the NDIA, not the AFP. That is not a criticism of either. It does mean that pointing a question at the police, as the committee did, will not reach most of the evidence base, and it puts more weight on the answers the agency itself gives on notice.

The next day, the sector argued about the word

On 31 July the witnesses were providers, researchers, health unions and disability organisations. Two exchanges are worth setting beside the police evidence.

Damian Bell, Group Chief Executive Officer of Vertaview Group, was asked about people not been treating the system with respect financially. He agreed there was a problem and then declined the framing:

I think people use the word ‘fraud’ very openly, and I don’t like to use the word ‘fraud’ as a blanket statement across a number of different things. I think you could probably label some of it ‘misuse’ or ‘use that wasn’t aligned to the purpose of the scheme originally’.

And Rob Oakley, Secretary of Capital Region Muscular Dystrophy, was asked what it would mean to be required to demonstrate he had tried other pathways before the scheme. His answer was that it would be humiliating, and that it is ridiculous to suggest that somehow or other I’m trying to defraud somebody by pretending to have a disease.

Clare Sullivan, Chief Executive Officer of MND Australia, put the mainstream sector position in one sentence: the organisation supports stronger safeguards against fraud, greater protections and accountability, but its concern is that the savings don’t come at the cost of delivering critical care.

Our view, labelled as such

The fraud the AFP and ACIC described is real, serious and, on the ACIC’s evidence, sometimes violent. Nothing here argues otherwise, and a reader who takes away “the fraud is exaggerated” has read this piece wrongly.

The gap is a different one. A Bill justified by fraud imposes its costs on providers, and those costs do not land evenly: a compliance obligation that a large provider absorbs into an existing back office can be the marginal cost that closes a small one. Deciding that trade-off well requires knowing where the fraud actually sits. On 30 July the Commonwealth’s police told the Senate that nobody has built the box to tick. That is a reasonable thing for an operational agency to say about its own case management system. It is a harder thing for a Parliament to legislate around, eight days before a report is due.

The second observation is smaller and more practical. If tier two is the vast majority of matters, then the questions that would settle this are ones only the NDIA can answer, and the committee has until 6 August to get them.

Primary sources

1. Senate Community Affairs Legislation Committee, Hansard transcript, 30 July 2026: the AFP and ACIC evidence, the exchange on large-versus-small provider data, the three-tier service-level agreement, and the Chair’s statement on questions on notice. Read 3 August 2026.
2. Senate Community Affairs Legislation Committee, Hansard transcript, 31 July 2026: the Vertaview, Capital Region Muscular Dystrophy and MND Australia evidence quoted above. Read 3 August 2026.
3. Senate Community Affairs Legislation Committee, Public hearings and transcripts: the hearing schedule, including the 6 August Perth hearing still to come, and the published programs listing every witness for each day. Read 3 August 2026.
4. SchemeBrief, One Bill, five peak bodies, four positions: the submission-stage positions this hearing evidence sits against.

Appeared before this inquiry, or read the transcript differently? Tell us and we will check it against the Hansard and log the outcome here.