SUPPLEMENTARY ATTACHMENT TO SUBMISSION ALREADY LODGED

Senate Community Affairs Legislation Committee

National Disability Insurance Scheme Amendment (Securing the NDS for Future Generations) Bill 2026

This supplementary attachment strengthens the governance, administrative law and public accountability submission already lodged with the Committee. It adds four components not included in the original: (1) a governance timeline tracing the trajectory from 2011 to projected 2028 commencement; (2) an administrative law chapter examining the Bill’s interaction with established principles of Australian administrative law; (3) a delegated legislation and parliamentary accountability section; and (4) a Committee Questions appendix providing senators and parliamentary staff with specific questions they may put to government witnesses.

Submitted by: Renee Fisher Independent Researcher and Legal Analyst May 2026

Supplementary Attachment — NDIS Amendment (Future Generations) Bill 2026 — Renee Fisher | Page 1

The Submission in One Sentence

The Bill does not primarily alter supports through individual decisions. It creates an administrative governance framework through which support levels, access thresholds, funding allocations, reassessment pathways and participant classifications can be adjusted systematically through rules, instruments, operational guidance and automated processes — without individual ministerial decisions and without being subject to individual merits review.

The submissions already lodged with the Committee demonstrate the Bill’s architecture in detail. This supplementary attachment adds the historical trajectory, the administrative law analysis, and the questions the Committee may wish to put to government witnesses.

The factual foundation for the core thesis is now fully public. Internal departmental modelling tabled in the Senate on 28 May 2026 confirms: 241,000 participants to be exited from the scheme in four years; peak participant numbers falling from a projected 944,000 to 598,000 by mid-2031; $37.8 billion in total savings targeted; and $13.2 billion of that coming from a single line — cuts to community participation budgets.¹ The Parliamentary Library’s Bills Digest describes the Bill as ’primarily aimed at reducing projected growth in NDIS expenditure and participant numbers.’²

These figures confirm that what is being debated is not an administrative efficiency measure. It is a structural contraction of the statutory boundary of the scheme — and the Bill provides the governance architecture to execute it.

Chapter 1 — What Changed Since 2013

The Bill is the current stage of a governance transformation that has been in development since at least 2019. Understanding the trajectory matters because each stage has narrowed participant rights, reduced discretionary space, and migrated operational authority downward from primary legislation. The Committee is asked to consider the Bill not in isolation, but as the point the trajectory has now reached.

Two observations from this timeline warrant the Committee’s attention.

First, the 2020 independent assessments proposal was defeated because it was visible — the assessment tools were known, debatable and contestable. The 2026 Bill achieves a materially similar functional result through a different mechanism. The specific thresholds, peer benchmarks, assessment methods and scoring logic that will determine eligibility for 241,000

• people are not in the Bill. They will be delivered later through NDS Rules and non-legislative incorporated documents. Parliament is being asked to authorise the framework before its most consequential operational settings have been determined.

• Second, the University of Melbourne institution that co-chaired the 2023 Review recommending the new assessment approach has subsequently received a contract to build, implement and conduct the validation research for the I-CAN assessment tool to be used under the new framework.⁴ The Committee may wish to examine what conflict of interest assessment was conducted in relation to this procurement and what independent validation arrangements exist.

The Key Point In 2020, the community could read the assessment tool and oppose it. In 2026, the assessment tool does not yet exist in legally binding form. The legislation creates the power. The tool arrives later through channels below full parliamentary scrutiny. That sequencing is the structural design, not a drafting oversight.

Chapter 2 — Administrative Law

Considerations

This chapter examines the Bill’s provisions against established principles of Australian administrative law. It is not a prediction of how courts will rule. It is an identification of the legal terrain courts will need to navigate — and, more immediately, grounds on which administrative decisions under the new governance framework may be challenged.

2.1 The Governance Framework: Where Power Resides

Before examining specific principles, it is useful to map where the operative authority under the Bill will reside. The following framework shows the chain from Parliament to participant outcome and identifies the accountability characteristics of each layer.

• Parliament (primary legislation) — publicly visible; subject to debate ↓
• NDIS Rules (delegated legislation) — subject to disallowance ↓
• Incorporated Documents (s 209(2AA)) — explicitly ‘not a legislative instrument’ ↓ [no disallowance; updated on NDIA website without parliamentary process] Standard Operating Procedures — internal instruments; limited public visibility ↓ Automated Decision Systems (Sch 3, Pt 2) — computer programs acting as CEO ↓ Participant Outcome

Democratic visibility and accountability diminish at each step down this framework. The most consequential operational decisions — what functional capacity score triggers exclusion, how peer benchmarks are constructed, what circumstances are deemed to exist or not — will be made at layers three and four, which carry the least accountability.

2.2 Fettering of Discretion

The common law principle against fettering of discretion requires that a statutory decision-maker must genuinely exercise their discretion in each individual case. A decision-maker who pre-commits to applying a fixed rule regardless of the circumstances presented fetters their discretion unlawfully.

The question the Bill raises is not whether any individual instrument is impermissible. It is whether the cumulative layering of thresholds, classification bands, SOP routing and automated processing leaves a delegate able to genuinely exercise the individualized discretion Parliament conferred. This question has particular force when the government’s own documentation describes the new planning process as one in which ’instead of reviewing individual supports in a plan, a review will look at the decision to accept the assessment as the basis for calculating the reasonable and necessary budget.’⁵

If the delegate’s role is to verify that the assessment was correctly accepted, not to apply the s 34 test to each support, the question is whether the s 34 discretion has been exercised at all — or whether it has been operationally transferred to the assessment instrument and budget method.

The Robodebt Structural Parallel

The Royal Commission into the Robodebt Scheme identified a deterrent architecture: participants bore the burden of disproving automated calculations; the challenge process used the same methodology that produced the original figure; the structural incentive was to accept the system outcome rather than risk a worse one. The replacement SNA mechanism in the Bill’s reassessment framework replicates this pattern. A participant challenging an assessment faces a replacement process using the same instrument, which ‘may increase, decrease, or not change the assessed support needs.’ That structural risk asymmetry discourages challenge. The Commission found this arrangement unlawful in the Robodebt context. The Committee should consider whether it is acceptable here.

2.3 Mandatory Relevant Considerations and the ICF Regression

A decision is invalid if it fails to take into account a consideration the statute makes mandatory. A decision-maker cannot be directed by policy or instrument to exclude a factor that the statute makes relevant.

The NDIS Act has always operated within the framework of the WHO International Classification of Functioning, Disability and Health (ICF) — a biopsychosocial model that treats disability as the interaction between impairment and environment. The ICF explicitly identifies environmental factors, participation barriers, assistive technology, and support networks as mandatory inputs to any assessment of functional capacity.

Section 9B introduces a definition that moves in the opposite direction — toward what disability scholars call the ICIDH model (the 1980 predecessor the ICF replaced), which locates disability entirely within the individual as a functional deficit. The new definition assesses capacity ‘without assistance from other people, assistive technology or modifications’ and ‘in a context that excludes, as far as possible, the impact of the person’s environmental and personal circumstances.’

The legal consequence is a potential conflict between the new s 9B definition and the Act’s existing objects and principles — which continue to require support for participation, choice and control, and social and economic engagement. Where the Act contains competing statutory signals, a court must attempt to reconcile them. The objects clause and general principles (s 4) may constrain the extent to which context can be excluded even under the new definition.

2.4 Procedural Fairness

Procedural fairness requires adequate notice of the case against a person and a meaningful opportunity to respond. Three provisions of the Bill create structural risks.

Provision Risk Nature

• s 50A — automatic plan renewal Funding levels adjusted by operation of law, without notice or opportunity to respond; explicitly declared not a reviewable decision Structural — no decision-point exists at which the participant can respond

• Schedule 3, Part 2 automated decisions Decisions generated through computer programs; the natural justice obligation cannot attach to an automated process Structural — accountability attaches to the decision-maker; where no human makes the decision, the obligation is difficult to satisfy

• s 48A — 90-day reassessment window Participant facing urgent support needs must wait 90 days for a decision; no interim support mechanism identified in the Bill Timeliness — procedural fairness includes timely process where delay causes serious harm

2.5 Merits Review Jurisdiction and the ART

The Administrative Review Tribunal’s statutory function is merits review — standing in the shoes of the decision-maker and making the decision afresh on the merits. The ART must be able to exercise the same discretion as the original decision-maker.

Where the decision-maker’s role has been reduced to verifying that a budget method algorithm ran correctly, the ART constrained to the same budget method may not be conducting merits review — it may be conducting algorithmic compliance checking. NDIS Rules cannot convert the ART’s merits review jurisdiction into a formula verification function; if the budget method purports to do this, it may be ultra vires the ART Act read with the NDIS Act.

The non-reviewability of s 50A automatic plan renewals and s 34A Ministerial support determinations removes entire classes of consequential outcomes from merits review entirely. A participant whose funding is reduced through a s 34A determination cannot challenge the quantum of the reduction before the ART — only the legality of the instrument. A participant locked into an underfunded automatic renewal has no reviewable decision to contest.

The Review Rights Question

If the s 34 reasonable and necessary test is no longer applied to individual supports; if the budget is calculated algorithmically; if renewals occur automatically by operation of law; and if class-level funding reductions are made by non-sunsetting ministerial instrument — what, precisely, remains available for a participant to challenge through the ART’s merits review jurisdiction?

2.6 Legal Unreasonableness and s 34A

A decision is legally unreasonable if it is so unreasonable that no reasonable decision-maker could have made it. Section 34A raises this question at a systemic level: the provision permits the Minister to fund supports at a level knowingly below their actual cost, where those supports have already been determined to be reasonable and necessary.

Historically, the NDIS logic was: determine reasonable and necessary supports, then fund them. Section 34A introduces a fiscal override that makes those two steps separable — and makes the second step discretionary. The legal question is whether there is a point at which the gap between determined need and actual funding becomes so disproportionate to the Act’s objects that it cannot be characterised as a legitimate exercise of a sustainability power. That question will eventually arise in proceedings. The Committee should consider whether Parliament intended to authorise structural underfunding of supports that have cleared the existing statutory threshold.

Chapter 3 — Parliamentary Accountability and Delegated Power

The concern addressed in this chapter is not that Parliament delegates power. Delegation is a necessary and routine feature of modern governance. The concern is the cumulative caracter and democratic visibility of what is being delegated in this Bill — and the specific mechanism in s 209(2AA) that explicitly removes a class of operational documents from parliamentary oversight.

3.1 The Scope of Delegation

The following categories of authority are delegated under the Bill to instruments below primary legislation:

The cumulative effect is that the most consequential operational decisions — those that determine whether a particular person can access the scheme and at what funding level — are concentrated in the layers with least democratic visibility.

3.2 Section 209(2AA) — A Dedicated Focus

Section 209(2AA) warrants specific examination because it is not a standard delegation provision. It is an explicit statutory declaration that a particular category of document — one that can contain scoring frameworks, assessment guidelines, classification criteria and operational manuals — will not be a legislative instrument. That declaration is deliberate. It has direct consequences for parliamentary oversight.

s 209(2AA)

| Parliamentary scrutiny | Standing Committee review | Minimal — website publication only | | Disallowance | Available — Parliament can revoke | Not available | | Sunsetting | Subject to sunsetting rules | Not subject to sunetting | | Registration | Required on Federal Register | Not required | | Public consultation | Typically required for significant instruments | Not required | | Amendment process | Parliamentary process | Updated by NDIA; website notification only |

The practical consequence of this architecture is that the rules most participants actually encounter — the scoring tools, the eligibility thresholds, the assessment manuals used in planning meetings — may live in documents that Parliament cannot disallow and that can be updated without formal process.

Most participants will never read the NDIS Act. They will experience the assessment tool, the planning guidance, the operational manual, the automated workflow. If those documents are where real access outcomes are determined, and if those documents are below parliamentary visibility, then democratic accountability for the NDIS’s most consequential functions has been formally removed by the Bill itself.

The Core Parliamentary Accountability Question

If the most consequential operational rules governing 760,000 people’s access to disability support are placed in documents that Parliament cannot disallow, that are not registered as legislative instruments, and that can be updated without formal process, what is the practical meaning of parliamentary oversight of the NDIS? This question is posed not rhetorically but analytically: the Committee should identify what accountability mechanisms remain operative after s 209(2AA) is enacted.

3.3 The Disability Discrimination Act Dimension

Under the Disability Discrimination Act 1992 (Cth), indirect discrimination occurs when a condition, requirement or practice is imposed on everyone but disproportionately disadvantages people with a particular disability, and the requirement is not reasonable in the circumstances. The hardest indirect discrimination question raised by the Bill is not whether any individual provision is facially discriminatory. It is whether the governance framework as a whole will produce predictable differential exclusion outcomes for particular disability cohorts.

Three specific risk vectors arise from the new functional capacity governance framework:

• The intrinsic ability assessment model (s 9B) may structurally privilege disabilities that are biomechanically stable, visibly physical, and legible to standardised scoring. Autism and psychosocial disabilities are characterised by fluctuating, context-dependent, and often invisible presentations. A standardised benchmark system assessing intrinsic capacity may systematically score these participants as higher capacity than their actual functional reality warrants.

• The appropriate treatment requirement (s 25A) may operate as a socioeconomic filter. A participant able to access private specialists exhausts treatment and is deemed permanently impaired; a participant on a multi-year public waitlist in a regional service desert remains technically “untreated.” The provision creates an access differential that correlates directly with income and geography.
• The direct impairment linkage requirement may partition the support needs of participants with complex comorbities in ways that fail to reflect integrated human functioning. Autism paired with mental health conditions, or acquired brain injury with chronic pain, produces support needs that cannot be cleanly traced to a single eligible impairment. The partitioning architecture creates “orphaned supports” — needs that exist but cannot be attributed to any single eligible pathway.

Chapter 4 — Evidence Register

The following register separates factual claims (grounded in primary sources) from analytical inferences (interpretive judgments drawing on those sources). Every factual claim is attributed to its source. This separation is intended to assist the Committee in distinguishing what the Bill and its supporting documents actually say from what this submission argues about their implications.

4.1 Fiscal and Statistical Claims — Primary Sources

4.2 Bill Text — Key Provisions

Explanatory Memorandum – Key Statements

Analytical Inferences – Confidence Assessment

toward the discredited 1980

description; Parliamentary Library and documentable ICIDH (functional deficit) aknowledgment of biosocial model model

The governance framework Government’s own NFP High — government creates cumulative fettering documentation states review documentation confirms the structure; legal consequence s 25A reverses Davis [2022] FCA 1002 Davis held financial circumstances relevant to ‘available’ treatment; s 25A(b) explicitly removes geographic/financial access from consideration High — direct provision comparison; EM confirms intent to reverse judicial expansion Automatic renewal and s 34A funding reductions may hollow merits review jurisdictional argument is sound; whether ART accepts it is unknown Moderate — jurisdictionalART constrained to same formula as delegate Community participation cuts ($13.2B) will generate cross-portfolio displacement confirmed; state foundational supports not yet operational generates cross-portfolio displacement unmeasured pending implementation Human rights literature on participation and isolation; no alternative delivery system confirmed; state foundational supports not yet operational

e

Appendix — Questions for Government Witnesses

The following questions are provided for the Committee’s use with government witnesses. They are organised by provision. Each is designed to elicit a factual, evidence-based response that either confirms or tests the analytical claims in this and the substantive submissions already lodged.

On Section 9B — Functional Capacity Assessment

  1. What specific assessment tool or tools will be used to operationalise the functional

capacity definition in s 9B? Has the tool been selected? If so, when will it be published for public scrutiny? 2. The Explanatory Memorandum states participants will be compared to their peers. What peer groups will be defined? By whom? Through what process? Will peer group definitions be published as legislative instruments subject to disallowance? 3. The Technical Advisory Group will advise on thresholds for ‘substantially reduced functional capacity.’ What is the TAG’s composition, appointment process and accountability framework? Will TAG recommendations be published before being incorporated into Rules or non-legislative documents? 4. Has the government assessed whether the intrinsic ability assessment model complies with the WHO International Classification of Functioning, Disability and Health, which the NDIS has historically applied? What independent validation has been conducted on the proposed assessment instrument for autistic participants, whose functional presentations are documented to be context-dependent, fluctuating and subject to masking? 5. Will the University of Melbourne — the institution whose co-chair participated in the Review recommending the assessment approach and which has received a contract to build and validate the I-CAN tool — have any role in validating the instrument’s reliability for NDIS eligibility purposes? What independent validation arrangement exists?

On Section 34A — Funding Reduction Power

  6. What actuarial modelling supports the funding percentage reduction levels the

government intends to apply under s 34A? Has that modelling been independently reviewed? Will it be published? 7. The Bill permits funding to be set below the actual cost of reasonable and necessary supports. How will the government ensure participants whose supports are structurally underfunded can supplement funding? Is it the government’s position that unmet need is an acceptable outcome of the new framework? 8. Section 34A determinations are exempt from sunsetting. What mechanism will Parliament use to review and revoke determinations if they are found to produce disproportionate harm? 9. Has the government modelled cross-portfolio liability displacement — specifically, whether savings from s 34A reductions will generate increased expenditure in health, housing, justice and income support systems?

On Section 25A — Appropriate Treatment

10. The Bill provides that treatment is appropriate even if inaccessible due to finances or
   geography. This directly reverses the Federal Court's interpretation in NDIA v Davis
    [2022] FCA 1002. What policy reason justifies reversing a judicial finding that
     specifically protected low-income and rural participants?
11. For participants in regional and remote areas where specialist services are
    unavailable: under the new framework, how will permanence be established? Is the
   government's position that these participants remain 'untreated' indefinitely?

On Section 209(2AA) — Non-Legislative Governance

12. Why should operational documents that directly determine participant eligibility —
    scoring frameworks, assessment guidelines, classification manuals — be exempt
   from parliamentary scrutiny and disallowance? What accountability mechanism
    replaces disallowance once these documents are published on the NDIA website?
13. What notification mechanism will ensure that participants, their advocates, and the
    public are informed when an incorporated document is updated in a way that
   changes eligibility outcomes?
14. If the most consequential rules governing access to disability support for 760,000
    Australians reside in documents Parliament cannot disallow, what is the practical
   meaning of parliamentary oversight of the NDIS?

On Automated Decision-Making (Schedule 3, Part 2)

15. Which specific decision types are intended to be fully automated? Will a public
     register of automated decision types be maintained?
16. What independent audit mechanism will be established to verify that automated
    decisions comply with procedural fairness, natural justice, and the duty to provide
   reasons?
17. The CEO substitution power — the ability to substitute a human decision where the
   automated outcome is 'not correct or not preferable' — under what criteria and
    timelines will this be exercised? Who monitors whether it is being used
    appropriately?

On Overall Architecture

18. The tabled modelling confirms 241,000 existing participants will exit the scheme by
   mid-2031. Which disability cohorts — by type, geography, age and income — are
    projected to absorb the largest proportions of that exit? Has the government
   conducted an indirect discrimination impact assessment under the Disability
    Discrimination Act 1992?
19. The Bill contains 240 amendment items. The most consequential operational details
— thresholds, assessment methods, peer benchmarks — are deferred to future
    instruments. Does the government consider Parliament has sufficient information to
   assess the Bill's impact on existing participants? If not, why should commencement
     of the access framework not be delayed until that information is available?

Supplementary Attachment — NDS Amendment (Future Generations) Bill 2026 — Renee Fisher

| Page 17

References

1 Jervis-Bardy D, ‘New NDIS eligibility rules will cut 241,000 participants from scheme in four years, documents reveal’, The Guardian (Australia), 28 May 2026, citing departmental modelling tabled in Senate. 2 Parliamentary Library, Bills Digest — NDIS Amendment (Securing the NDIS for Future Generations) Bill 2026. 3 NDIS Review, Working Better for People: Final Report of the Independent Review of the NDIS (December 2023). Professor Bruce Bonyhady AM was co-chair of the Review. 4 University of Melbourne/Melbourne Disability Institute contract for I-CAN tool build, implementation and validation research, 2024–2025. See also Australian Autism Alliance, statement on I-CAN validation concerns (2025). 5 Department of Health and Aged Care, Reviews and Appeals under the New Way of Planning, January 2026, p 2. 6 National Disability Insurance Scheme Amendment (Securing the NDIS for Future Generations) Bill 2026 (Cth). 7 Explanatory Memorandum, NDIS Amendment (Securing the NDIS for Future Generations) Bill 2026 (Cth). 8 National Disability Insurance Agency v Davis [2022] FCA 1002 (Mortimer J). 9 Mulligan v National Disability Insurance Agency [2015] FCA 544 (Mortimer J). 10 National Disability Insurance Scheme Act 2013 (Cth), ss 3, 4, 34. 11 Administrative Review Tribunal Act 2024 (Cth). 12 Disability Discrimination Act 1992 (Cth), s 6 (indirect discrimination). 13 Convention on the Rights of Persons with Disabilities, arts 3, 5, 19, 26. 14 World Health Organization, International Classification of Functioning, Disability and Health (ICF) (2001). 15 Royal Commission into the Robodebt Scheme, Final Report (July 2023), particularly Chapter 11 (deterrent architecture findings). 16 Legislation Act 2003 (Cth), Part 4 (disallowance), Part 6 (sunsetting). 17 Minister for Immigration and Citizenship v Li [2013] HCA 18 — legal unreasonableness standard.

Renee Fisher Independent Researcher and Legal Analyst — May 2026

Supplementary Attachment — NDIS Amendment (Future Generations) Bill 2026 — Renee Fisher | Page 18