NDIA Accountability Project

Financial modelling and forecasting of the scheme

‹ PrevPage 1 of 13 · Source p. 1Next ›

Inquiry into the Future of the NDIS

Submission 25 February 2022

Richard Madden AM, PSM, BSc, PhD, FIAA

Honorary Professor, University of Sydney

Introduction

This submission adds to my submission of 29 October 2021. Except where further addressed here, the points I made in that submission remain my view.

This submission relates to

Term of Reference (e): Financial and actuarial modelling and forecasting of the scheme

I completed ‘Comments on the NDIS 2020-21 Annual Financial Sustainability Report’ in December 2021. A copy of these Comments has been provided to the Committee. I will draw out and expand on some important matters raised in the Comments in this submission.

Since my Comments were completed, Taylor Fry’s ‘Review of NDIA actuarial forecast model and drivers of scheme costs’, completed in November 2021, has been released. I will highlight some important comments from their review.

Finally, I draw attention to the lack of consideration of the impact of mainstream services on the NDIS, despite this being part of the legislated mandate for the Annual Financial Sustainability Report.

Comments on NNDIS 2020-21 Annual Financial Sustainability Report

(AFSR)

The Summary contained in my December Comments was as follows:

   For the first time, the Annual NDIS Financial Sustainability Report, prepared by
    the NDIS Scheme Actuary, has been publicly released. This new transparency
   from the National Disability Insurance Agency is very welcome.

  The projections of NDIS costs to 2029-30 make sobering reading.

          Year                Participants         Total Payments ($ million)
          2020-21                 466,619                      23,347
          2024-25                 670,400                      41,373
          2029-30                 859,329                      59,284

   Key matters for discussion arising from the Report include:

       1. The sharp rise in the number of participants aged 0-14

      At 30 June 2021, 192,870 participants were aged 0-14, representing 42.9% of
        participants aged 0-64. This differs markedly from the 29.4% in the
       Productivity Commission’s 2011 costings.
        Exit rates are lower than expected. A large rise in participants aged 15-24 is
        projected in the years from 2025 (121,833) to 2030 (191,015); the proportion of
           all participants in this age range is projected to rise from 18.2% to 22.2%.

       2. The dramatic impact of participants aged 65 and over

     The number of participants aged 65 and over is projected to increase by
       267.8%, from 16,581 (30 June 2021) to 60,987 (30 June 2030), compared to a
      77.4% increase for participants aged 0-64. Over the same period, payments to
        participants aged 65 and over are projected to increase by 479.4% from $1.2
         billion to $7.1 billion, compared to a 135.9% increase for participants aged 0-
        64.
     The NDIS recommended by the Productivity Commission in 2011 covered
      people only up to age 64.

                                       3

Page 4

Projected payment growth for participants aged 0-64 is driven by participant numbers, not growth in real terms average payments.

The Report’s payment projections are shown only in current year dollars.

There are no real terms projections, using the dollars of a base year. In fact, real terms comparisons in the paper show that projected real terms payment growth for participants aged 0-64 is driven by the growth in participant numbers, not growth in average payments. This should be a key driver of attention for NDIS policy review.

The number of participants aged 15-64 at 30 June 2021 was 257,168, which is below the Productivity Commission estimate of 290,290. Overall then, NDIS costs for participants 15-64 are consistent with original expectations.

The continuing intake above expectations of new participants from regions where the NDIS has been in place for several years

Participant intake has not flattened out over time, despite the NDIS being in place in some areas since 2013. This underpins the projected ongoing increase in participant numbers up to 2030. There is no information on the payments going to these ‘late’ entrants compared to earlier entrants.

Conclusion

These matters highlight the need for more detailed analyses than is contained in the 2020-21 Annual Financial Sustainability Report, as a base for public policy discussions.

Expansion on December Comments

Each of these points is now addressed in more detail, expanding on material in my December Comments paper (references are to the Comments paper):

1. The sharp rise in the number of participants aged 0-14

The NDIS 2020-21 Annual Financial Sustainability report (AFSR) shows that the number of participants at 30 June 2021 aged 0-64 was 450,038 (Comments, Table 1). Of these, 192,870 were aged 0-14 (Table 3), representing 42.9% of participants aged 0-64. This differs markedly from the estimate used by the Productivity Commission in its 2011 costings: Table 16.2 of that Report shows 120,960 participants aged 0-14 out of 411,250 aged 0-64, or 29.4%.

Page 5

The number of participants aged 15–64 at 30 June 2021 was 257,168 (Table 1), which is below the Productivity Commission estimate of 290,290 (Productivity Commission 2011,

Chapter 16, Table 16.2).

The AFSR discusses the much lower than expected non-mortality exit rates (2020-21 AFSR Executive Summary and Figure 24). The 2011 Productivity Commission projected number of participants aged 0-14 was based on the 2009 ABS Survey of Disability, Ageing and Carers (SDAC). This cross-sectional survey showed a much larger number and percentage of people with severe or profound disability at ages 5-14 relative to the ages 15-24 decile. However, it could be expected that if a person aged under 15 became an NDIS participant, they would be reluctant to exit given the security of current and, importantly, future supports offered by the NDIS, so a lower than projected exit rate is not surprising.

SDAC results since 2009 show some increase in these numbers and percentages up to the 2018 SDAC. Notably, the percentage of the population aged 15–24 with severe or profound disability rose sharply from 2015 to 2018 (3.0% to 3.6% for males, 2.1% to 3.3% for females).

The projections show a large rise in participants aged 15–24 in the years from 2025 (121,833) to 2030 (191,015); the proportion of all participants in this age range rises from 18.2% to 22.2%. The Report comments:

  Young adults represent a growing proportion of the Scheme’s participant numbers as the
    children from the intake of prior projection years begin to age and transition into older age
   bands.

The fact that the number of people aged 0-14 is so much above the 2011 estimates, and that they are expected to lead to significant increases in the 15-24 age group, is a serious issue for the NDIS going forward.

The 2020-21 report contains a number of Recommendations about the future of the NDIS. Recommendation 9 is important here:

   The NDIA should reassess eligibility for participants who entered through early intervention as
    part of the normal course of business, whilst providing reassurance to those who exit the Scheme
    that it will be available to them in the future should their circumstances change.

2. The dramatic impact of participants aged 65 and over

The NDIS cost estimates prepared by the Productivity Commission in 2011 did not include people aged 65 and over (see Productivity Commission 2011, Chapter 3).

Page 6

However, the NDIS Act provided for participants who became participants before age 65 to remain in the NDIS after that age.

The 2020-21 Report projections separately identify the number of participants and the costs for people aged 65 and over. The State disability services systems which preceded the NDIS included relatively few people aged 65 and over. But the more generous provisions of the NDIS relative to the aged care system now give little incentive for NDIS participants to transfer to the aged care system when they reach the age of 65. Participant numbers and total costs are increasing far more quickly for people aged 65 and over than for people under the age of 65.

The number of participants aged 65 and over is projected to increase by 267.8%, from 16,581 at 30 June 2021 to 60,987 at 30 June 2030, compared to a 77.4% increase for participants aged 0-64 (Table 1). Over the same period, payments to participants aged 65 and over are projected to increase from $1.2 billion in 2020-21 to $7.1 billion in 2029-30, a 479.4% increase compared to a 135.9% increase for participants aged 0-64 (Table 2).

The increasing impact of people aged 65 and over is largely independent of the current decisions of the NDIA. Discussion about reasons for, and possible policy changes in response to, increasing costs in the NDIS should consider participants aged 0-64 separately from those aged 65 and over.

Drift from the aged care system into the NDIS poses a serious threat to the NDIS. The need for separate consideration of participants after the age of 65 is made more urgent by the developing campaign about inequities for people aged 65 and over with a disability depending on whether the disability was acquired after age 65 (not eligible for the NDIS) relative to those who are NDIS participants.

- Projected payment growth for participants aged 0-64 is driven by participant
  numbers, not growth in real terms average payments

Comparing 2029-2030 expenditure to 2020-21 expenditure is meaningful only if done in real terms. The 2020-21 AFSR does not include real terms estimates, although it does use percentages of GDP at different timepoints. The Commonwealth Treasury’s 2021 Intergenerational Report used real terms figures as well as GDP percentages when commenting on the NDIS.

Calculation of real terms estimates are shown in Table 5 of my Comments paper.

NDIS costs projection

NDIS costs in 2020-21 dollars are projected to increase by 78.1% over the 9 years from 2020-21 to 2029-30 (which is equivalent to an average real compound rate of 6.6%pa over the period). This is a most interesting result because the number of participants aged 0-64 is projected to increase at a similar rate, 77.4%, between 2020-21 and 2029-30 (Comments, Table 1). In other words, it is the growth in participant numbers that is driving the increase in real terms expenditure, not growth in average payments per participant.

This result demonstrates the need for real terms projections. The use of nominal terms estimates could lead to inappropriate comparisons when analysing the development of the NDIS.

4. The continuing intake above expectations of new participants from regions where the NDIS has been in place for several years

The AFSR reports on the continuing high incidence of new entrants in areas where the NDIS has been in place for several years (Figure 23 of the Report). One example is given (Report, P6):

The rate of new entrants to the Scheme continues to be high in geographical areas where the
Scheme has been operating for several years. As an example, the rate of new entrants in
  geographical areas that commenced in 2013 is 341 per 100,000 people, which is approximately
93% higher than the (previously) assumed rate of 177 per 100,000.

The NDIA has detailed information on those who have become participants in these areas (notably Hunter in NSW, Barwon in Victoria) at various dates since 2013, and the characteristics of people now joining the NDIS in those areas. Given the significant role that this continuing flow of these ‘late’ new entrants plays in the projections, much more analysis and discussion is needed of the characteristics of the new participants and their likely support needs compared to earlier entrants. In particular, are package sizes for those becoming participants after the NDIS has been in place for several years lower than those for earlier entrants?

In light of these ongoing rates of new participants in areas where the NDIS has been in place longest, explicit allowance for duration of the NDIS in the participant’s area in the modelling approach for participant numbers should be considered.

Taylor Fry Review

The Taylor Fry Review contains a wealth of discussion about a range of issues affecting projections of participant numbers and average costs. It is a welcome addition to knowledge about NDIS cost drivers and expenditure projections.

Matters addressed by Taylor Fry which have already been discussed are not repeated here.

Data provided for the Review

Section 1.7 sets out the wide range of data provided to Taylor Fry for their review:

  • De-identified unit record data pack which includes the following datasets, as at Jun-21:
    • Access requests: Information on plan approvals, access entry type, exits, etc.
    • Participant demographic: Information on participant’s age, gender, residence, activeness status, disability type, functional level, etc.
    • In-kind payments: Information on date and payment amounts
    • Payments: Information on support category, payment date, status and amount
    • Plan support: Information on plan effective dates, SIL indicator, budget amount by support category, etc.
    • Data dictionary defining the terms in the unit record data.
  • Annual snapshots of unit record patient demographic dataset from Jun-17 to Jun-21
  • AFSR model packs from Jun-16, Jun-18, Jun-19, Dec-19, Jun-20, Dec-20 and Jun-21 reviews which include:

    • Key assumptions, participant number and cost projection workbooks

    • Snapshot summaries of key information (in Dec-20 and Jun-21 AFSR packs only)

    • Scenario analysis workbooks (not in Jun-16 AFSR pack).

  • NDIA’s analysis on changes in level of functionality over time
  • Mapping between level of functionality and severity score by disability type
  • Breakdown of superimposed inflation from the 2018-19, 2019-20 and 2020-21 AFSR
  • Jun-21 snapshot with young people in residential aged care indicator flag
  • Summary data on participants who entered via s25, as at 30 June 2021
  • Assumed monthly seasonality factors for new participant entries and payments for 2021/22
  • Reference package cost mapping for FY 2020/21
  • NDIA’s report on Participant Level of Function and Primary Disability Type Change Analysis dated 30 September 2019
  • NDIA’s report on New Incidence Rates Analysis as at Dec-20 and Jun-21
  • NDIA’s report on Disability-Related Health Supports dated June 2020
  • NDIA’s report on NDIS implementation and Productivity Commission costing
  • NDIA’s report on Reference Package and Guided Planning Framework Review dated Apr-18
  • NDIA’s reports on analyses on exits, new incidences and payments as at Dec-19

AFSR Model Structure

It is assumed that confidentiality of participant records has been properly protected in the handing of data to Taylor Fry.

Detailed data and statistics on the characteristics of NDIS participants, their packages, supports used and their costs have not been available to stakeholders for analysis to inform their views on issues facing the NDIS.

The information base provided to Taylor Fry should be made widely available to all stakeholders (including researchers) for discussion and analysis of the NDIS.

It is arguable that much of the debate about the roll-out of the NDIS, its problems and its costs, could have been more manageable and less contentious if all stakeholders had access to the same database and reports as has been provided to Taylor Fry.

This argument for information transparency is not seeking to require the NDIA, or any relevant government agency, to make public internal working papers and communications.

AFSR Model Structure

The lengthy discussion of the AFSR Model Structure raises a range of important issues, which merit careful consideration. The following are analysed in some detail by Taylor Fry:

  • No explicit assumption of future movement between levels of functional ability
  • The model projects participant numbers independent of the prevalence rate it implies
  • The model is highly dependent on assumptions for ‘unanticipated superimposed inflation’.

There is comment on an alternative modelling approach which could be appropriate given the NDIA’s rich database on participants:

 > 'An alternative structure is a more person-centred model, where existing
  supports become the starting point for estimating what costs will be next year, at
   either an individual or cohort level (e.g. cohorts could be defined based on
 payment levels). Such an approach is adopted for many of the motor-injury
   lifetime care Schemes which were set up in parallel to the NDIS. This does not
   solve all issues; there are still uncertainties related to the distribution of new
   entrants and some aspects of cost escalation. But it gives a more direct way of
  understanding year-on-year increases unhindered by the current cohort
   definitions.'
  (Taylor Fry, P.14)

One part of the current model not discussed by Taylor Fry also merits review.

The model is based around ‘disability-severity’ bands. Taylor Fry’s Figure 3.3 shows a disability-severity matrix with 57 occupied cells based on 15 health conditions and 9 severity bands. While a health condition is necessary for NDIS eligibility, the use of health conditions for partitioning NDIS participants, rather than their level of functioning, is not in accord with the spirit of the NDIS. It suggests that people with the same functional needs could receive different overall resources from the NDIS because of their health conditions.

Future NDIS expenditure reductions due to early interventions and capacity building

Section 4(13)(c) of the NDIS Act provides that

    Reasonable and necessary supports for people with disability should…
      (c) develop and support the capacity of people with disability to undertake activities that enable
      them to participate in the mainstream community and in employment.

One of the aims of the NDIS was to provide supports to participants which will decrease their longer term need for supports. The Productivity Commission’s 2011 Report (Productivity Commission 2011, Volume 2 Chapter 16, P779) argued

     Early intervention by definition should result in cost-effective treatment which reduces costs in
    the long-term. Hence the $650 million in early intervention will reduce care and support
    requirements over the long-term. It is estimated that this reduction would be approximately
   $324 million

The 2020-21 AFSR did not include any cost offset as proposed by the Productivity Commission. Taylor Fry address the issue as follows:

  'Review of available evidence of early 'investments' reducing subsequent participant costs'
  Capacity building costs are 18% of total payments, with CB Daily Activities alone being 12%. These services are aimed at improving participants’ independence or allowing them to meet social, employment or other goals.

     It is recognised that it will take time to test whether these ambitions can be achieved, so currently the Scheme is paying for core services to meet existing levels of independence as well as capacity building, which is one reason why costs have increased rapidly over the past few years.

 …If the programs are successful, then this investment should lead to reduced core support costs for many participants and potentially more investment in similar programs.

…The projection model assumes growth in both core services and capacity building costs per participant. For the reasons stated above, if the NDIA has the proper controls in place then from around 2025 onwards when capacity building outcomes should be well understood the model is likely to overstate aggregate expenditure in this context.’ (Taylor Fry, P.35-36)

In light of this discussion, the NDIA should consider including a cost offset for early investments and capacity building in years up to 2029-30 in future Annual Financial Sustainability Reports, certainly for years after 2024-25. Their analysis should be informed by stakeholder input and the results published and included in future AFSRs.

                                    11

Supports outside the NDIS for people with disability (mainstream services)

Section 4(14) of the NDIS Act specifies

   People with disability should be supported to receive supports outside the National Disability
   Insurance Scheme, and be assisted to coordinate these supports with the supports provided
   under the National Disability Insurance Scheme.

Section 31 provides that

  The preparation, review and replacement of a participant’s plan, and the management of the
   funding for supports under a participant’s plan, should so far as reasonably practicable…

     (e) consider the availability to the participant of informal support and other support services
    generally available to any person in the community

The NDIS Act ((Section 180B(1) (a)(iii) & (b)) requires the Scheme Actuary, in preparing the Annual Financial Sustainability Report

    ‘to assess any trends in provision of supports to people with disability otherwise
  than through the National Disability Insurance Scheme …’

and ‘consider the causes of those… trends’

The 2020-21 AFSR contains no material on this part of the Annual Financial Sustainability Report’s mandate. The Taylor Fry Report also does not address this part of the AFSR mandate.

Given the projected growth in participant numbers, this omission is important. Mainstream services, including health, education and aged care, are budget limited, and under continual budget pressure in the face of high demand. It is not the role of the NDIS to ‘plug holes’ in mainstream services, which should be available and accessible to all Australians including people with disability (in line with the UN Convention on the Rights of Persons with Disabilities, of which Australia is a signatory).

The dramatic actual and projected increases in the number and cost of participants aged 0-14 make it vital to report on the supports provided to participants aged 0-14 by health and education services.

The projected increases in the number and costs of participants aged 65 and over, at a time of known high pressure on aged care services, make consideration of this area in future Annual Financial Sustainability Reports urgent.

Conclusion

The NDIS faces significant challenges:

  • Rapid growth in the number of participants 0-14
  • Low exit rates which are leading to significant growth in participants aged 15-24
  • Dramatic growth in participants and expenditure for those aged 65 and over.

In contrast, the number of participants aged 15-64 is within original estimates. Real terms expenditure for this age group is projected to grow in line with participant numbers.

The NDIS Annual Financial Sustainability Reports in future years need to be constructed so that attention is focused on areas facing growth pressure (such as those listed here), to ensure policy and administrative attention on NDIS sustainability is best focused.

As well, the Reports need to consider the potential future financial impacts of early interventions and capacity building.

The Reports also need a focus on the interaction of the NDIS and mainstream services, as specified in the NDIS Act.

                                     13