Australia's New National Disability Insurance Scheme: Implications for Policy and Practice

Abstract Australia is immersed in its largest reform of disability services in a generation – the staged rollout of the National Disability Insurance Scheme (NDIS). Enacted swiftly to capitalize on rare bipartisan political and public support, the $AU22billion scheme promises to design and deliver disability services differently, with emphasis on service user choice and control. However, the scheme’s rollout is outpacing the readiness of service users, service providers and the agency charged with implementing it to achieve its stated aims, threatening to derail it. Research suggests issues arising in the implementation of the scheme can be attributed in part to the design of the policy and, in part, to how it is translated into practice, both making scant reference to lessons from comparable reforms. Reflecting on the implications of these findings for policy and practice, we explore a range of challenges that arise when embarking on large scale reform in an environment of tight timelines, high expectations, extant policy silences, competing priorities, and jurisdictional ambiguities. This paper adds to the growing body of literature on the NDIS by bringing in the views of participants, and elaborating on implementation challenges posed by its governance structure, system architecture, and administrative structures.


Introduction
The National Disability Insurance Scheme (NDIS) mirrors shifts towards individualized funding for self-directed care in Western Europe, North America, and other OECD countries (Needham 2013). The scheme aims to increase both funding and access to services and support for people with permanent and significant disabilities and the control that people with disabilities have over the design and delivery of their care, and frames this not as welfare, but as insurance underpinned by actuarial analysis, economic modeling and human rights (Miller and Hayward 2017, 130;Productivity Commission 2011;Australian Government 2013, 4), It involves a shift from Australian governments block-funding disability services to a personalized service model, in which packages of funding from a single pool are allocated to eligible individuals, according to their level of need and self-defined goals, to purchase services and support from providers under a range of conditions. Implementing the scheme calls for actions and engagement across multiple organizational and individual domains, testing the capacity of government, its agents, public and private service providers, community partners, people with disabilities and their support networks to process information and to find new ways to interact (Carey, Kay, and Nevile 2017;Carey, Dickinson, and Olney 2019;Hill and Hupe 2009;Klijn and Koppenjan 2000;Rhodes 2007).
The NDIS has been scrutinized by researchers, policymakers and practitioners, service providers, service users, lawyers, journalists, and advocates since its inception, and its progress charted in academic research, stakeholder investigations and reports, government reviews, the media, and the courts. The record shows the scheme has made great strides in improving access to services and support for people with disabilities, but it risks derailing in the transition to full implementation (Productivity Commission 2017, 8-14). This can be attributed in part to the readiness and capacity of service users to understand and navigate the scheme; in part to the readiness and capacity of service providers to adapt and change in response to new funding arrangements and respond to consumer demand; and in part to the readiness and capacity of the government and its agents to build and monitor the scheme's architecture on a robust evidence base. The scheme was founded on promises of increased funding and autonomy for people with disabilities and assurances that its benefits would outweigh its additional costs (Productivity Commission 2011, 976). Yet a review of its costs by the Australian Government's independent economic research and advisory body in 2017 acknowledged that the scheme's financial sustainability hinges on (1) it being successful "on the balance of objective measures and projections of economic [and] social participation and independence, and on participants' views that they are getting enough money to buy enough high-quality goods and services to allow them reasonable access to life opportunitiesthat is, reasonable and necessary support"; and (2) that "contributors think that the cost is and will continue to be affordable, under control, represents value for money and, therefore, remain willing to contribute" (Productivity Commission 2017, 7). These are lofty aims.
In this paper, we draw on the academic and gray literature surrounding the NDIS to reflect on challenges for both policy and practice in rolling out a reform of this scale. We examine issues faced by practitioners in operationalizing the NDIS legislation, issues faced by service providers in entering or transitioning to a new market, and in particular, perceptions of service users concerning their experiences of the scheme. We aim to enhance the literature surrounding the NDIS by bringing in the views of participants, elaborating on implementation challenges posed by the scheme's governance structure, system architecture and administrative structures, and extracting lessons from the way the scheme is being implemented that can be applied in other policy contexts grappling with tight timelines, high expectations, competing priorities, and jurisdictional ambiguities.

Background of the reforms
The first object of the National Disability Insurance Scheme Act 2013 is to give effect to Australia's obligations under the Convention on the Rights of Persons with Disabilities (Australian Government 2013, 4;United Nations 2018). Parties to the Convention recognize the right of people with disabilities to "live in the community with choices equal to others" and undertake to "take effective and appropriate measures to facilitate full enjoyment by persons with disabilities of this right and their full inclusion and participation in the community" (United Nations 2006). The case for establishing the NDIS rested on evidence that Australian disability services were not up to this task. A review of the service system in 2011 found that disability services were underfunded, inflexible, fragmented and built around the needs of the system rather than those of individuals (Productivity Commission 2011). The funding and organizing of services were complex and inefficient, with gaps and overlap in state and territory, and federal responsibilities (Purcal, Fisher, and Laragy 2014), and the system's complexity was compounded by insurance-based funding of disability services and services funded privately as a result of public liability claims. Accordingly, access to services and support varied with individual circumstances, giving rise to inequity within the system. Studies also revealed persistent poor quality of life for Australians with disabilities, citing high levels of social exclusion and risk of violence, and low levels of income and Labor market participation (OECD 2009;Milner et al. 2014;Mithen et al. 2015;Kmjacki et al. 2016).
The push for a more equitable and personalized model of disability care and support was backed by a strategic and coordinated public campaign launched in 2011, designed "to appeal to Australians' sense of fairness" (Every Australian Counts 2016). Fronted by people living with disability and their families and carers, the campaign encompassed television and print advertising, lobbying local representatives, email, social media, public events, and rallies to garner support. Importantly, it was strengthened by an unprecedented alliance between the peak body for disability service providers, the peak body for advocacy organizations, and the peak body for family carers, groups that had historically been in conflict (Every Australian Counts 2016). The reform of disability services was framed as a human rights issue for people with disabilitya position no politician could reasonably or comfortably counter.
Under pressure to meet expectations raised by the campaign, the Australian Government developed a "highly ambitious" rollout schedule for a reform of its magnitude (Productivity Commission 2017, 10). Four trial sites were established when the NDIS was enacted in 2013, with a promise to roll the scheme out nationally by 2020 (NDIS 2017a), allowing little time to reflect on lessons from the trials and from comparable reforms (Duffy 2013, 10). Many people with disabilities have benefited from the reforms to date through shorter or no waiting lists for services, but data from trials reveals that swift and sweeping changes to disability services and funding in Australia aiming to give service users more choice and control over their care have not had the desired effect for all (Warr et al. 2017;Productivity Commission 2017, 76;Conifer 2017;Knause 2017). Other countries that have introduced comparable models of disability services have typically done so with attention to building and strengthening markets over longer timescales. For example, England created a market for disability services in the late 1980s and early 1990s as local governments divested themselves of disability services. Phased introduction of different options for individualizing funding for care services followed, starting with the introduction of direct payments in 1997, before personal budgets were introduced in 2008 with an expectation that these would become standard for all people with disabilities by 2015 (Glasby and Littlechild 2016). Although the English reform program was developed in a progressive manner (i.e. it did not set out initially to implement individual budgets), the length of time invested to restructure the market and prepare providers, policy makers, and service users was significant. Australia is designing and implementing an even more complex care system in under seven years. How the NDIS works.
As of February 2019, Australia has three policy instruments focused on enhancing the quality of life for people with disabilities and their families and carersthe National Disability Agreement (Productivity Commission 2019), the National Disability Strategy (Department of Social Services 2017) and the NDIS. Each is underpinned by different types of intergovernmental agreements, outcomes and reporting arrangements. The Productivity Commission has proposed new architecture for disability policy in Australia, intended to unify the various agreements, strategies, and policies relating to disability under a new, overarching National Disability Agreement, but it is yet to be adopted (Productivity Commission 2019, 8).
The National Disability Insurance Scheme Act 2013 sets up diffused governance arrangements for the NDIS, separating policy design and administration, as shown in Table 1.
Operational Guidelines for the NDIA, based on the Act and associated Rules (Australian Government 2018), guide "decision makers on how to apply the law in the context of the NDIS, and helps the NDIA to interpret its functions and powers when making decisions or recommendations affecting members of the public" (NDIS 2018d). The Guidelines are intended to be periodically reviewed and updated by the Minister. As noted earlier, the scheme's sustainability hinges on people with disabilities accessing universal or mainstream public services (Productivity Commission 2017). Access for people with disabilities to mainstream services, activities and programs is a shared responsibility between the NDIA and state, territory and Commonwealth government services, guided by principles outlining the roles and responsibilities of different sectors in that space (Council of Australian Governments 2015; NDIS 2018e). The NDIA and state and territory governments are supposed to work together to resolve issues relating to people with disabilities accessing vital supports on a case-by-case basis.
Australians under the age of 65 years with permanent and significant disabilities can apply to access funding under the scheme. If they meet the criteria, they attend an annual planning meeting with a planner employed by the NDIA or an agent of the NDIA to discuss their goals and the support they need to achieve those goals in the year ahead. The NDIA assesses whether the supports requested are reasonable and necessary, and whether they should be funded by the NDIS or if they are the responsibility of another service. Once a plan is approved, the NDIS participant can purchase services from a person, business or organization with their funding. They can choose to manage their own funding, appoint a third-party Plan Manager, or have their funding managed by the NDIA. If their funding is NDIA-managed, they can only purchase services from NDIS registered providers. If they are selfmanaging or using a Plan Manager, they can use NDIS registered providers and nonregistered providers.
As the NDIS rolls out, its complex governance and administrative arrangements, coupled with its geographic scale, quantum of funding and the complexity of the care market structure, its unclear interface with other services, and the broad range of needs and circumstances it must meet, are posing significant challenges to policymakers and practitioners Needham and Dickinson 2017;Malbon, Carey, and Dickinson 2017). Legal challenges to the assessment of eligibility for the scheme or assessment of reasonable and necessary supports are mounting (Morton 2018; O'Donovan 2018; Victorian Legal Aid 2018; Australasian Legal Information Institute 2019). Government inquiries into various aspects of the scheme are eliciting common themes around competing priorities, jurisdictional ambiguities and the importance of stewardship of the emerging marketsometimes referred to as quasi-market shaping or market interventions in public policy/administration literature to ensure equity of access to information, services and support under the scheme, particularly for citizens in remote locations or those with complex needs (for example Parliament of Australia 2019; ACT Legislative Assembly 2018). We anticipate ongoing adjustment of policy instruments and administrative arrangements in response to identified weaknesses in the scheme as its reach extends.

Research into the NDIS
The NDIS is a significant and high-profile policy reform initiative, so it is unsurprising that it is being closely watched. In addition to constant media attention, there is The scheme is administered by an independent National Disability Insurance Agency (NDIA), established under Commonwealth legislation and governed by a Board. The NDIA holds all funds contributed by the Commonwealth, States, and Territories in a single pool, manages scheme funds, administers access to the scheme, has decision-making powers over access, planning, information gathering and participants' nominees, and approves the payment of individualized support packages. 2 The NDIA Board is responsible for the performance of these functions and strategic direction of the Agency. The Board manages its costs and liabilities from year to year including through the development of a reserve and investment of funds. 3 The Board is advised by the NDIS Independent Advisory Council with up to 12 members (including a majority of people with a disability, at least two carers, at least one person with experience of disability in rural or remote areas and at least one service provider) and a principal member appointed by the Commonwealth Minister in consultation with all states and territories. Generally, all states and territories must support the appointment of other Advisory Council members. 4 The Standing Council on Disability Reforma Council of Australian Governments (COAG) Ministerial Council made up of Treasurers and Ministers responsible for disability from the Commonwealth and each State and Territorymakes decisions on NDIS policy. 5 The Commonwealth Minister is responsible for administering the NDIS Act and exercises statutory powers with the agreement of states and territories, including a power to make the NDIS Rules and direct the NDIA. 6 In a staged process between July 2018 and July 2020, an additional independent agencythe NDIS Quality and Safeguards Commissionis assuming responsibility for the registration and regulation of service providers in each state and territory.
increasing research interest in the scheme, with a range of papers, reports, and commentary on the NDIS being generated by different stakeholders (for example, David and West 2017;Dickinson, Needham, and Sullivan 2014;Productivity Commission 2017, 200;NDIS 2017c, 4;NDS 2017a;Every Australian Counts 2017). An audit of disability research in Australia in the time period 2000-2013 found that there was "not a critical mass of research on topics of priority to the National Disability Strategy, the National Disability Research and Development Agenda and the National Disability Insurance Scheme" (Llewellyn 2014, 6); an update just three years later found that "almost half of the policy papers addressed the National Disability Insurance Scheme (NDIS), with papers on operational issues and also more theoretical papers examining the broader social and health policy context in which the Scheme is operating" (Llewellyn 2017, 7). In just a few years, the field of disability has moved from being significantly under-researched to attracting greater interest and investment in research, although it remains to be seen whether this translates into better quality research and evidence generation. There is certainly room for more research focused on the knowledge and experiences of consumers and carers to critique and inform policy (Llewellyn 2017, 23).
The NDIS promises to "support a better life for hundreds of thousands of Australians with a significant and permanent disability" (NDIS 2017b). It aims to enhance independence, choice, and control by giving people with disabilities more choice about the type of help they receive, when they receive it, and who provides it. But a shift away from standardized services represents a shift of responsibility from the state to individuals, raising questions about how the scheme balances choice and control and safeguards for vulnerable citizens. Both practice and evidence demonstrate that individuals and families rarely live their lives in the way in which government services are organized . People living with a complex or chronic disability face challenges and issues that cross governance and administrative boundaries. As noted earlier, this is anticipated in policy, which specifies that access for NDIS participants to mainstream services is a shared responsibility between the NDIA and state, territory and Commonwealth government services, guided by principles outlining the roles and responsibilities of different service systems in that space (Council of Australian Governments 2015; NDIS 2018e). This is not playing out in practice. The 2017 review of the costs of the NDIS flagged that some supports were not being provided because the responsibilities of different levels of government to fund and deliver services were unclear, and called on governments to "set clearer boundaries at the operational level around 'who supplies what' to people with disability, and only withdraw services when continuity of service is assured (Productivity Commission 2017, 2). This reflects the "buck-passing" tendencies that Glendinning et al. (2011) observed in their research into individual budget pilot projects and the NHS in England, and goes some way towards explaining inconsistency in the scheme's outcomes. It is also reflected in research into participants' experiences of the NDIS (Warr et al. 2017).
When we began to examine the NDIS in its early stages, one of the gaps we identified in the evidence base was that there was more focus on aspects of the system and the impact of reform on government and the disability service system than on participants' experiences of the scheme. As noted above, this gap was also evident in the 2017 audit of disability research in Australia (Llewellyn 2017, 23). To balance this, in 2017 we conducted a study of the impact of NDIS reforms from the perspective of those using the scheme. Choice, control and the NDIS: Service Users' Perspectives on Having Choice and Control in the New National Disability Insurance Scheme was a participatory research project, intended to capture the experiences of people with disabilities, and family members providing care and support to people with disabilities, in one of the NDIS trial sites (Melbourne Social Equity Institute 2017). The findings were published in a community report (Warr et al. 2017). Interviews with forty-two participants of the scheme revealed mixed outcomes which fell under three themes: participants' experiences of choice and control in accessing services and resources; participants' views on the operation and efficiency of the NDIS; and insights into whether the NDIS was generating equitable outcomes (Warr et al. 2017). Key findings from that research that shed light on policy and practice are set out below, supported and supplemented by broader research and investigation where appropriate.

Participants' perceptions of choice and control in the scheme
In interviews, research participants talked about choice and control under the NDIS in different ways. Some focused on whether and how the NDIS was meeting their needs, while others talked about the extent to which they could choose the services and support they wanted. Being able to choose a service provider (Warr et al. 2017, 21), or choose when services were provided (Warr et al. 2017, 25) were welcome changes, despite limits on how funding could be spent: "They don't give you this $50,000 and say, 'Do what you want." (Warr et al. 2017, 22). Some participants felt they did not have the capacity to manage their own funding, or chose not to, and had engaged the NDIA or a Plan Manager to source and coordinate their services and shoulder the scheme's considerable administrative burden (Warr et al. 2017, 38).
Many participants described lack of clarity in the process of planning their care, particularly surrounding their level of choice and control and what constitutes "reasonable and necessary support" (Warr et al. 2017, 37). While the scheme promises participants choice and control, their preferences were subject to discretionary approval on the part of the scheme's care planners, who were described by many participants as lacking in-depth understanding of disability (Warr et al. 2017, 28-29, 33-37). Interviews also revealed frustration that boundaries between services participants wanted to streamline, like health, housing or education, remained pronounced (Warr et al. 2017, 40-41). This was particularly evident in discussion with parents of young children who were keen to integrate support for their children seamlessly into mainstream activities: I am really worried about it. I have heard a lot of things about how therapists cannot go into the classroom and things like that. Really worried about how that is going to impact his learning, basically, and how that is going to work. It has impacted us already. Before he was accepted into kinder, there was no day care centre that basically wanted us. They turned us away (Warr et al. 2017, 40). Several participants said they were unable to find and access services approved in their plan (Warr et al. 2017, 42, 45). This emerged as a significant concern in negotiating annual plans, which involve forecasting a person's needs over the next 12 months, where participants cited examples of requests for funding being declined on the basis that funds previously allocated for similar items had not been spent. Underutilization of agreed services, equipment and support was particularly evident among participants living in regional areas or those needing highly specialized services and supportsmarkets of limited appeal to providers of goods and services seeking economies of scale. However, the supply side of the equation was not the only issue. Our research also revealed that some participants were ill-equipped to push for new services or to utilize different services, either because they could not access information about their entitlements in a form that suited their needs and circumstances or because they risked being labeled as "difficult" (Warr et al. 2017, 49). Furthermore, if their health and wellbeing fluctuated over the life of their plan, as is common among people with both physical and psychosocial disabilities, they only needed some services, equipment, and support on an ad hoc basis. This scenario generated anxiety among some participants about the future consequences of not having access to the required services.
Whatever their circumstances, most participants described some gap between their expectations about choice and control, and the reality of navigating the NDIS. For those who had campaigned for reform in the lead-up to the NDIS, this was disappointing: I reckon we were at a meeting every fortnight about NDIS and NDIA and what it was going to be … with our campaign red t-shirts on, banging things and marching up and down saying, 'This is what we want!' And then it slightly changed from you will get everything and then it was, what is reasonable and appropriate and within reason? All those words suddenly started appearing (Warr et al. 2017, 50).
The planning process is the introduction for people with disabilities into a system they are likely to interact with for their whole life. Early damage to the relationship between people with disabilities and the NDIS, resulting in lack of trust and doubt about the scheme's capacity, will be very difficult to overcome and could have farreaching unintended consequences.

Participants' perceptions of the operation and efficiency of the service system
Several participants said the scheme's administrative burden outweighed its benefits. Discussion about the scheme's administrative burden was often linked to its ICT system, but participants raised numerous communication-related issues, including difficulties in obtaining reliable and accurate information (Warr et al. 2017, 29), communicating with the NDIS when issues arose (Warr et al. 2017, 37) and using the online portal to view plans, keep track of budgets, request payments and manage services (Warr et al. 2017, 37-38).
Some participants reported that the NDIS offered a clearer and less complicated pathway for integrating services than they had previously experienced. However, many others claimed that the scheme had created a new complex system with numerous chances for miscommunication and/or breakdown in continuity of care. Annual planning reviews required extensive preparation and expert evidence, especially for participants opting to self-manage their funds. The complexity of the service system and the need to synthesize complex information about administrative procedures and service providers meant that many participantsparticularly if they did not have access to informal sources of support from families and carersstruggled to access, process and apply this information. Participants in these situations had little sense of available options (Warr et al. 2017, 35-36).
The administrative burden surfaced again here, appearing to counter the scheme's aim to encourage unpaid carers to return to work. Participants we interviewed who had opted to self-manage did so because they felt sufficiently motivated and competent to assemble and manage a suitable portfolio of services for a family member. Most were parents of young children with disabilities. While they valued the opportunity to tailor a package of services and support that best meet the needs of their child, the opportunity cost was high: Working is something I really want to get back into and I will never be able to do the amount of hours a week that I did before. It is not something you can just sit down and the computer say on a Monday. I have thought this through really carefully. Could I just do three hours on a Monday of NDIS admin, meaning contacting all these people and what is next on the thing. Could I do it like that? No. Because it is picking him up one day at two o'clock from school because this particular person cannot see him at 3:30 or four. They can only see him at 2:15 and they are only available on a Friday. I have to fit in with them in order to make sure he gets his thing. (interview transcript #3) All participants accepted the need to account for the expenditure of public funding, but most reported that accountability requirements were disproportionately expensive and time-consuming for small items (Warr et al. 2017, 42-44). Participants' insights suggest that it may be more efficient, both for participants and the NDIS, if requests for funds under a certain amount, or requests for equipment where the need is clearly self-evident, could be approved without involving intermediary professionals. Building scope to respond to unanticipated contingencies into planning processes would also enhance overall efficiency without compromising the integrity of the scheme.

Equity
Participants' stories of general difficulties they experienced in navigating the NDIS system highlighted the ways in which factors that drive inequalities, such as age, gender, socioeconomic status, residential location, and household structure, operated as fault lines in the NDIS system by further constraining the potential for choice and control. These issues were rarely directly addressed in participants' accounts, but surfaced in discussions of related issues. One participant commented that the concept of choice is more familiar to individuals encouraged and accustomed to making decisions based on their personal preferences: A lot of families of people with disabilities didn't really seem to know actually what they want to do because they're not practiced at making choices, or making real choices.
They're practiced at making choices from what's available, but not necessarily saying, 'I don't like anything that you've got available for me, this is what I want 'cause that's what will make a difference to my quality of life' (Warr et al. 2017, 46) More broadly, participants who could draw on professional experience which gave them an understanding of the logic of meetings, preparing funding proposals and liaising with professionals were more confident and assertive in their interactions with the NDIS, compared to participants who had previously had limited exposure to these kinds of processes. Access to computers, the internet, and other material and digital resources enabled some participants to research and identify available opportunities, while others were excluded. "If you did not know the jargon you virtually got zilch," one participant said (Warr et al. 2017, 27). Questions of equity and fairness in the NDIS raise many practical and ethical issues that are largely unaddressed in the scheme's design.
Gaps between the scheme's promises and participants' experiences appear to stem primarily from inconsistent interpretation of "choice and control" and "reasonable and necessary support" in care planning and service delivery (NDIS 2017d;Productivity Commission 2017, 181;United Nations 2006 Article 19). The NDIS promises participants choice and control in planning their care, but their options for support are bounded by cost-benefit analysis and discretionary decision-making on the part of the scheme's care planners. Of particular interest from an implementation standpoint is the scheme's firm stance that it will not fund support that should be offered to people with disability by mainstream public services "even if the system response does not provide it" (NDIS 2017b, 3). The 2017 review of the scheme's costs confirmed that "[n]ot stepping in to replace other systems is a critical principle of the NDIS to ensure people with disability continue to be serviced by mainstream systems and the NDIS is financially sustainable" (Productivity Commission 2017, 3). Yet variations in the funding and organization of non-NDIS support across jurisdictions means there is no nationally consistent basis on which the NDIS can engage with other systems.
This challenge is not unique to the NDIS, but it has significant ramifications in this case. The scheme's emphasis on "mainstream interface" and the 2017 review of its costs flag a need for it to work across service boundaries to achieve its goals (Productivity Commission 2017, 2). To date, this has played out as a delineation of fiscal responsibility. Concerns have also been raised about the NDIA's low level of sector engagement, particularly in relation to the agency understanding complexities involved in the provision of services at the front line, unrealistic pricing and the level of red tape (NDS 2016, 18;NDS 2017b;NDS 2018). A lingering challenge for Australian federal and state governments is clarifying "who supplies what" to people with disability and balancing stakeholder expectations and available resources. It is clear that promises made in the lead up to reform of disability services extend beyond the scope of the scheme .

Implications for policy and practice
Many Australians with disabilities are faring better under the NDIS than under previous arrangements for disability care and support. However, inconsistent practice in care planning in the scheme is problematic. The funding structure and processes for identifying individuals' needs are complex and, to some extent, this is to be expected in a scheme of this size servicing people with a very broad range of needs and personal circumstances living in diverse communities and locations. However, there is clearly potential to simplify administrative systems, clarify aims and objectives of the scheme to promote consistency in decision-making when allocating resources and increase flexibility in response to the varying capacities and resources available to service users to comply with key administrative requirements, which in turn will build trust in the scheme. More investment in training and equipping NDIS staff for their roles, and in preparing and supporting service providers to transition from block funding to a contestable, client-driven market, may ease the concerns of service users about the capacity of the system to meet their needs, as well as reduce stress among care planners and service providers who are likely to be a target of participants' frustration with processes and access to services. The scheme's significant administrative requirements have placed additional pressure on people with disabilities and households that are already stretched in supporting family members with disabilities, and effectively exclude those without capacity or assistance to meet these requirements. Furthermore, approval processes for equipment that is clearly warranted and in some cases, inexpensive, not only eat into individual funding packages but put pressure on scarce professional services such as physiotherapists and occupational therapists in rural and regional locations. This suggests the rate of implementation should be slowed to ensure that service users and service providers are prepared and equipped to interact effectively and efficiently in the new regime.
In theory, the NDIS offers people with disability choice and control over their care by offering them access to a range of services and resources. In practice, services and resources are not consistently available to participants, and participants' capacity to exercise choice even when services and resources are available is variable. Key elements of self-directed support for NDIS participants revealed in our 2017 study were: being able to envisage and articulate their goals and what they needed to achieve them; having their vision and the support they wanted to be accepted by their NDIS planner as "reasonable and necessary"; and being able to access approved services or support when they wanted them (Warr et al. 2017). The scheme's success will hinge on layered stewardship, or market management, of the service system , as described in Table 2.
The 2017 review of NDIS costs found that in the trial phase, the NDIA "focused too much on quantity (meeting participant intake estimates) and not enough on quality (planning processes), supporting infrastructure and market development" (Productivity Commission 2017, 2). There has been encouraging progress since, with a range of pathway and market improvements being tested in practice, encompassing participant pathways, provider pathways, information about the assistive technology market, and changes to the NDIS online portal (NDIS 2018a; NDIS 2018b). In addition, the 2013 NDIS Act has been amended to establish an independent NDIS Quality and Safeguards Commission to oversee the delivery of supports and services under the scheme, to support participants to exercise choice and control, to ensure appropriate safeguards are in place for NDIS supports, and to establish expectations for providers to deliver quality support from 2018 (Department of Social Services 2017; Australian Government 2017). The government is heeding warnings that it must balance participant intake, the quality of plans, participant outcomes, and financial sustainability to achieve the aims of the NDIS, but it still faces significant challenges in relation to stakeholders' expectations.

Conclusion
In reflecting on research (and other) findings into the NDIS to date, we bear in mind that it is an ambitious social policy initiative. Criticisms or critiques of implementation of the NDIS must be considered within the parameters of the accelerated timescales within which the scheme has been developed, the huge scale of reform involved and the sorry state of many disability services prior to its introduction. The scheme's framework was broadly conceived and it was acknowledged from the outset that challenges would arise in implementation that would be addressed through a continuous improvement process enshrined in the Act (Australian Government 2013, 6). The grassroots movement that propelled it into being had high aspirations and expectations that it would deliver much-needed resources and support to some of Australia's most disadvantaged citizens. Having moved swiftly from being a broad set of aims and principles to being implemented in diverse communities, it is Supply of services and support to meet demand.

Supplement markets to address gaps
Market -Individual -Government/Agency readiness (crossover) Quality assurance and safeguards. Monitor service quality Actively encourage diffusion of best practice unsurprising that the scheme is stretching the capacity of government, its agents, public and private service providers, community partners, people with disabilities and their support networks. The key risk for the Australian Government lies in failing to address this early, before the scheme is irredeemably tainted. The vision for the NDIS, in some respects, outstrips what has been achieved in public sector markets internationally to date. This challenge is compounded by the breadth of the expectations, needs, and circumstances of the scheme's participantsas well as those of people with disabilities deemed ineligible for the schemeand persistent administrative boundaries between the NDIS and other public services that people with disabilities wish to draw together for seamless support. There is no blueprint for policymakers to build upon in designing a functional self-regulating market for disability care and support in Australia (Public Service Research Group 2017) and with governance of the NDIS diffused across multiple entities, responsibility for its stewardship remains unclear (Australian National Audit Office 2016). Until this is addressed, there will be ongoing challenges for those charged with ensuring that the NDIS fulfills the social contract between government and citizens with disability.

Disclosure statement
No potential conflict of interest was reported by the authors.