The Untold Truth The Untold Truth
Australia

The NDIS Cuts: Where Does the Money Actually Go?

The government says it will save $14.4 billion by cutting 160,000 people from the NDIS. But the replacement programs cost more than the savings. The modelling is unreleased. And nobody has explained what happens to the next person who needs help.

TU

Staff Writer

24 April 2026 • 11 min read

Live Investigation

The government says it will save $14.4 billion from the NDIS. It plans to cut 160,000 people from the scheme. But the replacement programs cost at least $10.5 billion. Some estimates put it above $16 billion. That is more than the savings. The modelling behind the cuts has not been released. The government says the scheme will shrink to 600,000 people by 2030. Independent analysts say the math does not add up. New entrant rates would need to drop sharply. About 60,000 First Nations people use the NDIS. Many live in remote areas with no other services. The government has not said how the cuts affect them. Disability does not pause for budget cycles. Demand keeps growing whether the scheme has a cap or not. The questions are straightforward. Where does the money go? Who loses access? And what replaces it?

Yesterday we published the numbers. 160,000 Australians removed from the NDIS. Plans cut by 16 percent. The government tightening participant eligibility while providers who overcharged the scheme face registration as their penalty.

Butler said this is about providers, not people. The numbers tell a more complicated story.

$14.4BHeadline NDIS savings claimed
160,000People the government says will be removed
$10.5B+Cost of replacement programs

Now the questions get harder. The $14.4 billion in headline savings has a problem. Several.


The modelling nobody can see

Butler stood at the Press Club and said the scheme will shrink to 600,000 participants by 2030. Down from 761,442 right now. He cited modelling.

The modelling has not been released. The Guardian reported on 22 April that Butler referenced “unreleased modelling” for the 600,000 figure. It cannot be independently verified. (Source: The Guardian Australia, 22 April 2026)

What we do know: about 19,000 new participants join every quarter (annualised, roughly 76,000 per year). Over five years to 2030, that is 380,000 new entrants. Start at 761,000. Add 380,000. Subtract 160,000 removals. You get 981,000. Not 600,000.

The math only works if new entrant rates drop sharply, or exit rates spike, or the 160,000 figure means something different to what it sounds like.

Turns out it might. Instagram account @the_yarn__ made the point the day of the announcement. “Labor’s projection of 160,000 fewer NDIS participants by 2030 reflects a reduction against earlier growth forecasts, not a reduction in total.” Not 160,000 people being removed. The gap between where the scheme was heading and where the government now says it will land. (Source: Instagram @the_yarn__, 22 April 2026)

A meaningful distinction. One the government has not made clearly.

The NDIA Annual Financial Sustainability Report breaks participants into “Projection Groups” based on new entrant rates and exit rates. That report exists. The detailed projections underpinning the 600,000 target have not been made public. (Source: NDIA AFSR, dataresearch.ndis.gov.au)

First structural question. If the modelling exists, release it. If it does not exist, say so. An accountability figure built on unreleased data is not accountability. It is a press release.


The blurry line for new applicants

What happens to the person who develops a disability next year? Or in 2028? Or 2030?

The government plans new eligibility criteria from 1 January 2028 for new entrants. Butler confirmed this at the Press Club. (Source: health.gov.au, Q&A with Minister Butler, 22 April 2026) What those criteria look like in practice is unclear. The government has committed to a “needs assessment” tool. The NDIS Review recommended “foundational supports” outside the scheme for people who miss the threshold. In theory, people who do not qualify for the NDIS get help elsewhere.

In practice, nobody has shown where “elsewhere” is. Or what it costs.

The numbers are worth considering. In 2022, 5.5 million Australians had disability. 21.4 percent of the population. In 2018, it was 4.4 million. A 25 percent increase in four years. (Source: ABS Survey of Disability, Ageing and Carers, 2022)

Autism diagnoses went from 205,200 in 2018 to 290,900 in 2022. A 41.8 percent increase. A peer-reviewed study published in December 2025 found the NDIS itself drove a 32 percent rise in reported autism prevalence. People who were always there. They just got diagnosed because a system finally existed to help them. (Source: ABS Autism in Australia 2022; Science Direct, December 2025)

Disability does not pause for budget cycles. People do not stop acquiring disabilities because the scheme has a growth target. The question is not whether the NDIS costs too much. Maybe it does. The question is whether the savings plan accounts for the fact that demand is structural. It does not care about the fiscal calendar.

Grattan Institute, an independent policy think tank, said in June 2025 that capping growth at 8 percent “will not be enough to make the scheme sustainable.” The government is now targeting below that. (Source: Grattan Institute, “Saving the NDIS,” June 2025)


The money nobody accounts for

The headline is $14.4 billion in savings. At the same time, the government is building replacement programs outside the NDIS. “Foundational supports.” Separate funding. Separate bucket.

December 2023. Albanese struck a $10.5 billion deal with states and territories to split the cost of non-NDIS disability services, in return for GST funding. (Source: The Guardian, 6 December 2023)

ABC 730, the night of Butler’s Press Club speech. Butler mentioned another $6 billion. “A couple of years ago, the National Cabinet agreed another $6 billion to be allocated… foundational supports.” (Source: health.gov.au, Butler ABC 730 transcript, 22 April 2026)

At minimum $10.5 billion in replacement funding, potentially $16.5 billion or more. The NDIS savings figure is $14.4 billion.

The replacement programs cost more than the savings.

The arithmetic is straightforward. NDIS saves $14.4 billion. Foundational supports cost at minimum $10.5 billion. The net position is not a saving. It is a cost transfer.

ABC ran a headline that day: “Labor’s NDIS savings plan hinges on state-run supports.” (Source: ABC News, 22 April 2026) The plan depends on states picking up the people the Commonwealth removes. The states signed up in principle. But the total cost of those supports has not been published against the savings figure in a single document anywhere we could find.

Which brings us to the question that matters most. Is the $14.4 billion a gross figure being presented as a net saving? Does it deduct the cost of replacing the NDIS for the people being removed? Or is money going out of one pocket without being shown going into the other?

The National Agreement on Foundational Supports is on the federal financial relations website. It says the Commonwealth and states “jointly fund and deliver additional supports.” (Source: federalfinancialrelations.gov.au) What it does not say, anywhere, is what the total replacement cost is when set against the NDIS savings.

The arithmetic is straightforward. NDIS saves $14.4 billion. Foundational supports cost at minimum $10.5 billion, potentially more. The net position is not a saving. It is a cost transfer. From the Commonwealth to the states. From a national scheme to a patchwork of state programs that do not exist at scale yet.


The people nobody is counting

The government publishes quarterly data on NDIS participants who identify as First Nations. What it does not publish is how the planned 160,000 reductions break down by remoteness or Indigenous status.

60,529 First Nations people were NDIS participants as at 30 June 2025. 8.2 percent of all participants. Indigenous Australians are 1.5 times more likely to have a disability than non-Indigenous Australians. (Source: AIHW Indigenous Health Performance Measure 1.14; ARK Services)

First Nations NDIS participants are already 28 percent less likely to receive care. Not our analysis. The Disability Royal Commission said that. (Source: Disability Royal Commission media release, 26 June 2023)

Remote areas: 36 percent of NDIS participants are Indigenous. Very Remote: 73 percent. (Source: NDIA data, dataresearch.ndis.gov.au)

19 March 2026. The NDIA released a “Statement of Commitment to stronger action on Closing the Gap for First Nations people with disability.” A month later, the cuts. (Source: ndis.gov.au, 19 March 2026)

Today. 23 April 2026. The First Peoples Disability Network called for no Indigenous participants to be transitioned off the NDIS. They want cultural safety built into the new assessment tool “from day one.” (Source: SBS NITV, 23 April 2026; Croakey Health Media, 23 April 2026)

Croakey Health Media reported last year that the changes were “likely to push many rural providers over the edge, to the detriment of NDIS participants and rural and remote communities.” (Source: Croakey, June 2025)

The Disability Advocacy Network Australia put out a joint statement. “People with disability, their families and the organisations that represent them are all worried.” (Source: dana.org.au)

All on the public record. None of it addressed in the government’s savings announcement.

The government will not say how many of the 160,000 reductions come from remote communities. Will not say how many are Indigenous. Will not say whether the cuts fall proportionally or disproportionately on the people with the least alternative support.

If you live in Sydney and get cut, there are other services. Not enough, but some. If you live in a very remote community where the NDIS is the only formal disability support that has ever existed and you get cut, there is nothing. The foundational supports that are supposed to replace the scheme in those communities do not exist yet.


The pattern

Four questions. Same answer.

What modelling supports the 600,000 target? Unreleased.

Does the $14.4 billion account for replacement costs? Not published.

Can the growth cap hold? Grattan says no.

Are the cuts hitting remote and Indigenous communities hardest? No data.

Headline savings. No modelling. No offset costs. No demographic reality. No distributional impact. Every question met with unreleased data or absent data.

The government built a $50 billion scheme without adequate controls. Their own admission. Now they are selling the fix without showing the math.


The other side

State governments already deliver many disability services through health, education and community programs. The foundational supports framework is partly about clarifying which level of government is responsible for what, rather than creating entirely new services.

The NDIS was never designed to be the only disability support system. The original Productivity Commission report envisioned a tiered model where the NDIS covered the most significant disabilities, with mainstream services handling the rest. Over time, the scheme became the default for a wider range of needs than originally planned.

Some current NDIS participants may be better served by general health and community services, particularly those whose primary needs are medical rather than disability-specific. The review found cases where people were placed on the NDIS because no other system was available, not because the NDIS was the right fit.

The $10.5 billion state deal includes GST funding adjustments. It is not a pure cost transfer from the Commonwealth to states. The GST component means states receive additional revenue to support the services they are being asked to provide.

Growth from $4 billion to $50 billion over a decade is faster than any other Australian government program. The trajectory was on track to reach an estimated $125 billion per year by 2034 if unchecked. No government, of any political stripe, has sustained that kind of growth in a single program without intervention.


The way forward

Release the modelling publicly. If the 600,000 target is built on sound projections, the data should withstand scrutiny.

Publish a replacement cost analysis against the savings figure. Australians deserve to see the net position, not just the gross cut.

Publish the demographic impact of the cuts by remoteness and Indigenous status. If the data shows the cuts fall proportionally, release it. If it shows they fall disproportionately, that needs to be addressed before the cuts take effect.

Ensure foundational supports are operational before moving participants off the NDIS. Removing people from a working system before the replacement exists is not reform. It is a gap.

Commission an independent review of the new assessment tool by the disability sector. The tool that decides who qualifies and who does not should not be designed and evaluated solely by the agency implementing the cuts.


Sources

Related Investigations

NDIS and Public Service Cuts