Public administration reform for Aboriginal affairs: An institutionalist analysis

Persistent underperformance of public policy and program implementation in Aboriginal affairs is widely recognised. We analysed the results of two case studies of attempted reforms in public administration of Aboriginal primary health care in the Northern Territory, using a framework based on the institutionalist and systemic racism litera-tures, with the aim of better understanding the sources of implementation failure. Implementation of the agreed reforms was unsuccessful. Contributing factors were as follows: strong recognition of the need for change was not sustained; the seeds of change, present in the form of alternative practices, were not built on; there was a notable absence of sustained political/bureaucratic authorisation; and, interacting with all of these, systemic racism had important consequences and implications. Our framework was useful for making sense of the results. It is clear that reforms in Aboriginal affairs will require government authorities to engage with organisations and communities. We conclude that there are four requirements for improved implementation success: clear recognition of the need for change in ‘business as usual’; sustainable commitment and authorisation; the building of alternative structures and methods to enable effective power sharing (consistent with the requirements of parliamentary democ-racy); and addressing the impact of systemic racism on decision-making, relationships, and risk management.


INTRODUCTION
Implementation of public policies and programs in Australian Aboriginal affairs is famously challenging, and instances of failure abound (see, e.g. the historical analysis of Aboriginal policy conducted by the Northern Territory Royal Commission into the Protection and Detention of Children; Commonwealth of Australia, 2017). Despite repeated calls for a different way of working together (Banks, 2009, pp. 14-15;Commonwealth of Australia, 2017, pp. 250-256;Department of Finance and Deregulation, 2010, p. 13), it seems that most reform attempts fall foul of the 'business as usual' problem.
There seems to be a repeating pattern of persistence with business as usual, with governments setting program goals and priorities in capital cities, followed by failures of engagement and/or relevance at local and regional levels, leading to poor program outcomes. In this corrosive process, community representatives and organisations see lack of commitment and respect, whereas bureaucrats and politicians see a different and more frustrating kind of implementation challenge than they encounter working with and through the mainstream community. There is a sense of circularity in cause and effect, which can be simply expressed in these terms: the lack of [political] [bureaucratic] [professional] commitment makes it hard to get anything done properly; and because it is hard to get anything done, there is a lack of commitment.
In this paper, we report an attempt in the primary health care (PHC) field to get beneath this conundrum. Specifically, we sought to understand why and how there is both a lack of commitment and a failure to accept that working with Aboriginal community representatives and organisations cannot be the same as contracting with mainstream NGOs for the delivery of services. This is an important goal precisely because failure is so frequent in spite of genuine attempts and agreements to do things differently.
We suggest that an institutionalist analysis, which emphasises established formal and informal patterns of interaction, provides a useful way to understand the mismatches in institutional settings and habits (as between government officers and authorities on the one hand, and Aboriginal communities and organisations on the other), and thus to identify some of the main requirements for success in the reform process itself.

CONTEXT
PHC policy and practice is one of the most successful areas of Aboriginal affairs. Long-standing policy commitments by Australian governments (Commonwealth of Australia, 2013;National Aboriginal and Torres Strait Islander Health Council, 2003) have been enacted in continuing investment in Aboriginal Community Controlled Health Organisations (ACCHOs), inter alia. But attempts to consolidate that commitment and develop a comprehensive nation-wide system of PHC for Aboriginal people centred on the ACCHO sector have repeatedly stalled. Centrally driven vertical programs and the use of competitive tendering have been preferred by the Commonwealth government (Dwyer, Lavoie, O'Donnell, Marlina, & Sullivan, 2011;Moran, Porter, & Curth-Ribb, 2014). Jurisdictional governments, which carry primary responsibility for the public health care system, seem unable to implement the needed structural reform in service systems (in this case, the transfer of appropriate PHC delivery from government health authorities to the ACCHO sector).
The sector provides effective PHC (Mackay, Boxall, & Partel, 2014; Thompson et al., 2013) to a significant proportion of the Aboriginal population in Australia (Australian Institute of Health and Welfare, 2016, p. 7;National Health and Hospitals Reform Commission, 2009, p. 87), and functions as an important part of the health system, working with mainstream healthcare providers. In some communities, they remain the only PHC provider available. In others, they are alternatives to mainstream services and preferred for their ability to provide culturally safe care and to advocate on behalf of patients within mainstream services (Gomersall et al., 2017). Relationships between the sector and governments are of long standing and relatively robust. There are many strengths to build on.
These achievements have been made in the context of widespread systemic racism towards Aboriginal people (Durey, Thompson, & Wood, 2012), and the unfinished business of coming to terms with dispossession and colonisation (Sullivan, 2011). In the absence of constitutional reform, public administration becomes the conduit and interface through which Aboriginal self-determination and decision-making in the policy process must occur (Davis, 2018). For policies and programs to be implemented with Aboriginal communities, decision-making and governance need to be understood with reference to dispossession, the ongoing process of colonisation (Davis, 2007;Sullivan, 2011), and the need for some form of constitutional reform and inclusion (Referendum Council, 2017). These underlying factors contribute to the apparent inability of public administrators to engage effectively with Aboriginal communities despite consistent acknowledgement of the need to do so.
The fundamental challenges that governments and communities face in finding effective ways to develop and implement mutually acceptable policies and programs in most fields also hold in PHC. For some years, governments and the sector have seemed to be stuck in an ambivalent stasis. The parties have different institutional settings and norms but need to work together in their respective roles as funder and provider of essential PHC. There is enduring high policy support for the sector, and fairly stable funding arrangements, developed through decades of both strong advocacy and collaboration. But PHC coverage is patchy, with some communities seriously under-served and without reasonable access to care when they need it (Australian Institute of Health and Welfare, 2015). Governments are neither enabling growth nor finding effective alternative ways to deliver essential PHC for communities, although there have been some attempts to bypass the sector, by moving funds to other non-government or private sector providers, in recent years (Henderson, 2015). Government reluctance to enable the growth of the sector in order to achieve better coverage has a direct impact on health outcomes for Aboriginal and Torres Strait Islander people who lack the ready access to essential PHC generally available in Australia. This paper analyses an attempt to change the relationship between Australian government health authorities in the Northern Territory and the ACCHO sector. The intent was to move away from treating the ACCHO sector as just another service provider to be funded in order to deliver services as determined by governments. The agreements articulated in Pathways to Community Control (Northern Territory Aboriginal Health Forum, 2009) (henceforth: Pathways) committed both levels of government and the sector (representing the health interests of Aboriginal communities) to implementing some fundamental shifts in authority, roles, relationships, funding, and accountability that would enable more community control over PHC governance and service delivery.
The planned changes, summarised in Table 1, had the potential to demonstrate a different way of doing business for both governments and the sector. They envisaged a transfer of responsibility for PHC delivery in Aboriginal communities from the NT health authority to regional community-owned organisations. There was to be a continuing strong role for the tri-partite NT Aboriginal Health Forum (henceforth 'Forum'), through which both national and NT government health authorities and the peak body for the sector, the Aboriginal Medical Services Association NT (AMSANT), negotiated continuing development and policy for Aboriginal PHC and for the sector.  These planned changes represented a shift to community control of PHC delivery, and network governance (Osborne, 2010) of the PHC system. That is although the governance of ACCHOs remains under community control, the broader PHC system needs to be planned, coordinated, and held accountable as a network among government and non-government bodies, including ACCHOs. Pathways expressed the directions and scope of reforms at a high level, and work commenced immediately after its release to develop more detailed plans for implementation. Table 2 summarises the main elements of the reform processes, with the location of primary responsibility for the work program for each element highlighted by shading.
Significant investments were made in developing detailed plans and proposals, and in shared governance of the reform process, but implementation was mostly unsuccessful, with one clinic being transferred during the period of the study, and no regional proposals being accepted by government health authorities.
Having observed and analysed the reform program at both jurisdictional and regional level, we then sought to make sense of what happened using an institutionalist lens.

THEORETICAL FRAMEWORK
Our starting point is that the institutional shift envisaged in Pathways entails a transformation from New Public Management (NPM) hierarchical governance, exemplified in the centrally driven approach to classical contracting for service delivery and accountability, to a form of network governance of Indigenous health services in the Northern Territory. We regard this as fundamentally a task of institutional change, and draw on institutionalist and public management literature to identify the specific conditions that are required for such change to occur.

Institutionalist theories
Under the broad banner of 'new public governance', many scholars have argued for a change from practices shaped by NPM to more collaborative, networked institutions of governance, particularly in domains of publicly funded social services (Osborne, 2010, Van Dooren & Hoffman, 2018. This argument has been echoed and amplified in work on governance of Indigenous services, by the present authors and others . Pathologies of NPM-inspired, 'classical' contracting include fragmentation of service delivery, high transaction costs for contracting and accountability processes, and the hard-wiring of low-trust relationships between funders and providers, and between competing providers (Dwyer, Boulton, Lavoie, Tenbensel, & Cumming, 2014;Lavoie, Boulton, & Dwyer, 2010;Smith & Smyth, 2010;. However, such institutional change is profoundly challenging for a number of reasons. Firstly, in Parliamentary democracies such as Australia, although network governance may have advantages for the practical management and effectiveness of complex service systems, such arrangements can never completely replace traditional Westminster frameworks of accountability. So the question of compatibility between hierarchical and network governance becomes central. As Klijn and Skelcher (2007) have argued, network governance may not fit neatly with pre-existing hierarchical forms. If the boundaries are not clear, or the incompatibilities serious, there are likely to be significant tensions between network and hierarchical governance.
Secondly, a significant amount of 'institutional work' is required to design and test changes such as this. As Richard Shaw (2013) argues, the implications for institutional design have been underestimated in NPG/public value literature, and new forms need to be imagined and built. Shaw suggests the following prescription, which fits with many elements of the proposed NT arrangements: …combining policy and delivery responsibilities in nondepartmental organizations which are closer to the communities they serve. Such public entities would be operationally autonomous from the political centre (but held accountable through, say, a fully or partially politically appointed board), and thus more flexible, nimble and responsive to local aspirations than central departments are considered to be (Shaw, 2013, p. 488).
Shaw and others (Cloutier, Denis, Langley, & Lamothe, 2016;Lawrence & Suddaby, 2006) also argue that the required institutional work involves tackling many issues of authorisation and accountability.
Next, we turn specifically to institutionalist literature to explore where the energy required for such change emanates from. As a broad approach, institutionalism focuses on the formal and informal rules of interaction that develop as ingrained habits as much as, or more than, the product of 'conscious' design. March and Olsen define institutions as the enduring collection(s) of rules and organized practices, embedded in structures of meaning and resources that are relatively invariant in the face of individual turnover and changing external circumstances (March & Olsen, 1989).
Their definition emphasises the 'stickiness' of institutions (Pierson, 2000). However, since around 2000, there has been a marked turn within institutionalist literature to understanding processes of institutional change (Hacker, 2004;Mahoney & Thelen, 2010). Although it used to be commonplace to argue that institutional change occurs predominantly due to exogenous shocks, there is now considerable attention given to processes of endogenous change -where the energy for change comes from within (Zietsma, Groenewegen, Logue, & Hinings, 2017).
One prominent author (Crouch, 2005) has theorised the nature of endogenous institutional change, making two crucial contributions. Firstly, he treats modes of governance (hierarchies, markets, networks, etc.) as combinations of more fine-grained institutional characteristics. Any change from hierarchical to network governance, therefore, requires multiple significant changes of habitual practice, of which three are particularly important: (i) From 'signalling' -where the content of behaviour that complies with institutional rules is pre-established (given), to 'dialogue' which allows for negotiation and exchange through communication.
(ii) From 'verticality' where communication implies a single centre of authority to 'horizontality' which develops rules for facilitating lateral communication.
(iii) From 'hard enforcement' based on formal compliance mechanisms (e.g. contractual penalties) to 'soft enforcement' in which mutual obligations within the service system discipline the behaviour of the parties. (Crouch characterises this contrast in terms of 'strong' and 'weak' enforcement, but in many service systems the weakness of market forces, and the interdependence and cultural strength of some networks make that distinction unreliable).
Crouch's second contribution is the idea that the seeds of institutional change need to come from within (as well as without). That is to say, within any regime dominated by particular institutional routines (e.g. hierarchy), there will be actors (institutional entrepreneurs) who will have developed alternative practices (e.g. network). These alternative and subordinate practices may eventually supplant or transform dominant institutions, particularly when coupled with external shocks, and/or with a realisation that existing institutional practices are sub-optimal or counter-productive. Crouch's framework has been successfully tested for its applicability to reforms in PHC (Tenbensel, 2018).

Systemic racism
In the administration of Aboriginal health policy and programs, systemic racism is an important characteristic of institutions. Systemic (or institutional) racism, as opposed to interpersonal racism, is the built-in kind of discrimination that functions without individual actors in the system having to do anything other than the normal (Jones, 2003). To put it more formally, institutional racism is 'those established laws, customs and practices which systematically reflect and produce racial inequalities…if racist consequences accrue…the institution is racist' (Jones, 1996, p. 438) whether or not individuals have racist intentions. There is good evidence of systemic racism in Australian health care, as seen in differential access and outcomes for Aboriginal patients (reviewed in Bourke, Marrie, & Marrie, 2018).
In the U.S. institutionalist literature, systemic racism has been explored by authors such as Sheingate (2014), who has argued that meso-level policy and administrative settings in the U.S. welfare system embody practices of 'neoliberal paternalism' that have clear racialised implications in jurisdictions with higher concentrations of African-American citizens. In the Australian context, this is perhaps best exemplified in the way public administration of Aboriginal affairs relies on centrally defined, tightly targeted programs and contracts that seem to multiply at the community level. Moran (2018) cites the example of a community in Western Australia where there were 90 services, programs, or projects in operation for a population of 360 people. The goals, scope, and methods to be used in these (mostly) government-funded interventions are almost all defined in distant capital cities. Local communities are likely to remain without any effective power in shaping this picture, and unsurprisingly, are mostly disengaged.
The program and contracting approach also contribute to the related problem of excessive use of red tape and compliance requirements. This model of contracting with Aboriginal organisationswhile familiar to other non-government service providers -is based on specific perceptions of risk in funding Aboriginal organisations, which are assumed to lack governance strength and operational capacity. This pattern of administration meets the definition of systemic racism, as it has a differential application and impact in Aboriginal communities and organisations.
We suggest that there is a kind of 'elective affinity ' (McKinnon, 2010) between NPM-based approaches to contracting with the ACCHO sector and the systemic racism of Australian society. Meanwhile the alternative framework that inspires and guides the Aboriginal health sector, including its focus on self-determination, inevitably conflicts with the mainstream NPM-inspired thinking and practices of public administration. The general lack of mutual trust (Reconciliation Australia, 2017, pp. 10-12) or shared sense of purpose is reinforced, and in turn there is a negative effect on the development and implementation of policy improvements and on working relationships. This is truly a wicked policy problem (Hisschemöller & Hoppe, 1995), and these well-known factors drag on both meaningful cross-cultural engagement and implementation success.

Requirements for successful reform in this context
From this discussion, we can identify four key conditions necessary for the type and scale of institutional change implicit in Pathways. Firstly, following Crouch, institutional change requires recognition 'from within' of weaknesses and problematic aspects of the dominant governance arrangements. This means that policy actors within government need to see that things need to change.
Secondly, the seeds of institutional change need to have already developed as alternative practices that find a niche within dominant institutions and serve as a template for institutional change from within. Thirdly, institutional change of this nature inevitably requires continuity of political authorisation from ministers and senior public officials. Because of the time required to redesign institutions, and the likelihood of turnover in political and senior management leadership, sustained political authorisation is challenging to achieve and maintain, and may require the setting up of alternative structures designed for this purpose, as suggested by Shaw (2013).
Fourth, measures are required to counteract the pervasive undermining effect of systemic racism. It needs to be acknowledged and addressed, and its impacts on working relationships within deliberative and coordinating structures need to be managed.

METHODS
With our industry partners, our team documented the Pathways reform attempts in the Northern Territory using two nested case studies, one at jurisdiction level  and the other at the level of a region (East Arnhem) .
We entered into an agreement with Forum, representing both levels of government and the ACCHO sector in the NT. We observed the processes of the reforms over 3 years (2011)(2012)(2013)(2014), analysing minutes of meetings of Forum and its various working groups, as well as policy and other working documents. We conducted formal interviews with 28 stakeholders and participated in meetings and other informal discussions. Interviews were recorded, transcribed, and analysed thematically. Text was initially coded using categories based on the major elements of the reforms and refined during the analysis using an inductive approach. The case study methods and results are reported in more detail in the original case study reports .
In this paper, we have conducted a secondary analysis of the Pathways case studies, using the analytical framework described above. We synthesised and interpreted the results in order to answer these questions: To what extent does our framework explain implementation failure? And could this framework be used as guidance for improving the success of reforms that require governments to work with Aboriginal organisations and communities?

RESULTS: THE PLAN, AND WHAT HAPPENED
Pathways (2009) was endorsed by all three parties after many years of negotiation. It laid out a plan for transferring government-run clinics in Aboriginal communities across the NT to regional community governance, along a continuum of increasing levels of community control according to community readiness or capacity. The fact that Forum members were successful in this negotiation was an achievement in itself, building on Forum's track record of success in improving and expanding PHC in NT Aboriginal communities. Over many years, the partners had developed a rational planning and evidence-based approach, and a strong capacity to work together and to try new ways of doing things (Allen and Clarke, 2011; Devitt et al., 2015).
There was a trade-off at the heart of this plan -in return for community control, communities would be required to develop PHC at a regional rather than discrete community level (there were planned to be about 14 regions). For communities, regionalisation brought the requirement to work across language and cultural differences, as well as sizeable areas. For some ACCHOs, it would mean trading local independence for more stable funding and expanded roles.
The advantage for government was that this reduced the potential number of governing bodies that would need to be established and then demonstrate capacity, and then be funded and held accountable. However, government health departments would be required to meet the challenge of working differently with the sector, both in PHC delivery and in overall planning and resource allocation (including through bundled regional allocation of funding, and funds pooling across vertical programs). That is health authorities would be required to move from a relationship based, at least formally, in arms-length contracts for discrete packages of centrally defined activities, to working with Aboriginal community organisations as providers of essential care for regions. This implied some power sharing -for example health authorities would have reduced power to unilaterally shift resources between vertical programs. For the NT government, there was another major challenge arising from the requirement to transfer existing services, clinic staff, and resources.
Importantly for all parties, regionalisation would enable some small clinics and organisations to achieve the critical mass needed for effective PHC delivery, and the new arrangements would enhance the capacity of the service system to allocate resources to high priority health problems and opportunities at local and regional levels.
The Pathways document focused strongly at the community and service delivery level -that is on the question of how is this going to work in practice? It paid little attention to the necessary changes in funding, regulation, the role of NT Health, and overall governance of the PHC system in the NT (J Moffet, CEO of NT Health, personal communication 12 February 2013).
Work began in 2010. A central planning unit was funded and established within AMSANT, and a tri-partite committee reporting to Forum took responsibility for the planning phase. Regional steering committees were established in four regions, charged with developing plans for the transfer of service delivery and for demonstrating community capability. Forum's attempts to address some of the system (as opposed to service delivery) requirements were largely unsuccessful, as both government departments took a 'wait and see' approach. For example no technical work was undertaken on the parameters and methods of funds pooling.
The deadline for submission of regional proposals was 30 June 2012, and in our second case study, the East Arnhem Steering Committee met this timeline. A great deal of development, consultation (Christie, Greatorex, van Weeren, & Cathcart, 2011), and planning had been done, in a fertile field because the communities of this region had previously determined that regional development in all fields of activity was their goal (Miwatj Health, 2013).
But by July 2012, there had been a change of government in the NT and of government leadership in the Commonwealth, and a major mainstream health system reform process (Council of Australian Governments, 2011) had resulted in the loss of key senior public servants who were members of Forum. Momentum for the Pathways reforms was steadily disappearing. In response to the East Arnhem regional plan, requests for changes were made by senior government officers, changes that would have undone carefully negotiated arrangements between local agencies and communities, and that were not essential to meeting the policy and operational requirements of the Pathways reform.
Further correspondence was entered into, meetings were held, Ministerial promises were made, and one NT government clinic was actually transferred. But in the end, no final response to the regionalisation proposal was ever made by governments. The situation in other regions was different, but the outcome was largely the same -no regional proposals were accepted, and the goal of transferring responsibility for the governance and management of PHC for Aboriginal people in the NT to communities and ACCHOs, with all of the system and funding changes such a transfer would require, was not achieved. Rather, the Pathways project experienced a largely unexplained passage into neglect and inaction that seems to be typical when endorsed policies or programs can neither be successfully completed nor publicly acknowledged to have failed.
Activity has in fact continued at both regional and jurisdictional levels, due to efforts by the sector and the NT health department (Cork-King, personal communication 8 December 2017). In East Arnhem, further clinic-by-clinic transfers have happened (Miwatj Health, 2018). But the Pathways attempt at a systematic policy-driven process of transfer ended with a whimper. Although there was some regional progress (as evidenced by the East Arnhem proposal), virtually no progress had been made at any stage on planning or developing the system-level public administration elements of reform.

ANALYSIS
The case studies demonstrate the extreme difficulty for public administrators to find feasible ways to move from a relationship formally defined and operated as funder-to-contractor, towards one of networked arrangements where the separate organisations shared some power and decision-making for the regional care system as well as for jurisdiction-level coordination and governance tasks.
We suggest that these case studies are not just another sad story of poor execution leading to implementation failure. Rather they exemplify a widely acknowledged pattern in Aboriginal affairs (see, e.g. Banks, 2009) that has deeper roots and needs to be understood as an institutional feature of Australian public administration.
Applying the four required conditions of our institutionalist theoretical framework to the Pathways reform attempt, we found some features that met the requirements for establishing a new institutional fit and initially facilitated progress, but were not sustained, and others that were not present or attempted. They are strong initial recognition of the need for change, founded on many years of policy work and relationship building; the seeds of change in the form of alternative practices; and a notable T A B L E 3 Conditions for institutional change and the Pathways experience

Condition for institutional change Pathways experience
Recognition 'from within' of weaknesses and problematic aspects of the dominant governance arrangements.
Strong recognition by senior public servants and ACCHO leaders of the need for change to improve health care delivery, and supportive policy was in place.
Seeds of institutional change already developed as alternative practices that find a niche within dominant institutions and serve as a template for institutional change from within.
Forum had a strong track record in collaborative planning and policy development that had enabled success in previous projects.
Continuity of political authorisation from ministers and senior public officials. Due to long timelines, may require alternative structures for the purpose.
Political authorisation secured but lost in changes of elected governments and high turnover of senior public servants. Forum constituted as deliberative not operational/coordinating body and no alternative was developed.
Measures to counteract the pervasive undermining impacts of institutional racism on working relationships need to be embedded.
Institutional racism recognised on all sides but not addressed in management or deliberative settings. Perception of high risk to government and to careers of senior personnel was a barrier. Community leaders felt that their tasks, and contributions, were undervalued.
absence of sustained political/bureaucratic authorisation. Interacting with all of these, systemic racism had important consequences and implications. These features are mapped to the theoretical framework in Table 3 and discussed below.

Recognition of the need for change
Forum representatives recognised that existing arrangements were not consistent with established policy nor with enabling the health gain that effective PHC can offer to Aboriginal communities, and together they created the opportunity that was crystallised in the commitment to Pathways. However, government recognition of the need for change, and its necessary extent, appears to have been limited to a particular group of senior public servants experienced in Aboriginal health in the NT. This fragile foundation was undermined when high staff turnover, brought on by broader national health system reforms commencing in 2011, removed key members of this group. The required recognition, and commitment to the pursuit of change, were not sustained.

Alternative practices seeded, but failed to thrive
For any network governance approach to be workable for governments, suitable public administration modalities are required in order to enact the required switches. Firstly, these are from signalling and vertical communication to dialogue and horizontal communication (as is implied in the over-used notion of 'partnership' and as required for the functioning of networks). For example centrally defined program and contracting approaches are perhaps the most important enactment of signalling and vertical communication in this context. Neither are likely to be abandoned by governments, but rather a shift from multiple 'vertical program' contracts to fewer, integrated contracts, supplemented by networking arrangements, is needed. The third is to shift towards 'soft' enforcement (although contractual and network accountabilities would likely co-exist). This is in itself difficult but not impossible, and governments routinely use dialogue and horizontal communication in some contexts, not generally including Aboriginal affairs (e.g. in relations between public hospital boards and health ministers). Forum played the critical role in designing, initiating, and seeking to implement the planned reforms. It provided a space in which the actors could develop methods of governance that had network characteristics -that is participants had already established a niche in which they had the capacity to 'switch' to practices of horizontal communication, dialogue, and the use of persuasion rather than the hard threat of enforcement. Forum had thus proven its ability to negotiate some successful changes to funding and governance (such as through the Primary Health Care Access Program; Rosewarne & Boffa, 2004). This did, for a time at least, come close to what Crouch (2005) suggests is required as an endogenous force for institutional change within the dominant institutions. In the event, this force was not effectively maintained for long enough and may have been insufficient to weather the inevitable setbacks experienced in any implementation process, particularly in the context of Aboriginal affairs.

Need for sustained authorisation and alternative structures
There were initial and continuing problems with authorisation, in spite of dual ministerial sign-off on the Pathways document. This high-level commitment was not matched with secure structures and processes for authorising the needed changes. The allocation of implementation responsibility to Forum was well-intentioned (including for sharing power) but the required operational authority for this otherwise deliberative body was not negotiated. Forum members were not empowered to make decisions that would be binding on all parties, and Forum was thus unable to act effectively to maintain momentum and resolve problems in the implementation process. Although Forum served well as the site for the development of alternative practices, it was not suitably configured to function as an alternative ongoing governance structure for implementation. The failure to provide for implementation capability in this way was a critical contribution to project failure and is sometimes described as the classic approach to transfer: 'chucking it over the fence' without thinking through the authority and implementation structures required.
The more difficult and technical requirement was to reconcile network governance with hierarchical authority in the context of parliamentary democracy and electoral accountability (Klijn & Skelcher, 2007).
The development of such arrangements is a central dilemma for governments and Aboriginal polities in the attempt to negotiate and enact policies and programs that are mutually acceptable and more effective. The possibility for regional boards to fulfil a regional network governance function was contemplated in Pathways (and precursors were established in the form of regional clinical networks). But the potential for Forum (with appropriate changes to role, membership and structure) or an alternative body to take on a jurisdiction-level network governance function for the PHC system was not explicitly contemplated or discussed. The absence of active work within the Commonwealth and NT health authorities on required changes in their administrative arrangements to enable implementation of Pathways (such as the development of regional funding methods) was also a clear indication of failure to engage with the development of alternative institutional forms and practices at system (or jurisdictional) level.
Thus, the work on Pathways did not ever reach the point of contemplating the question of arrangements to ensure that a network governance body could operate separate from the major government agencies, while also ensuring appropriate authorisation and accountability, as suggested by Klijn and Skelcher (2007) and Shaw (2013). Proposals for such a body at the national level, made by the sector and by the National Health and Hospitals Reform Commission (National Health and Hospitals Reform Commission, 2009, pp. 73-74), did not gain traction with Commonwealth health officials.
The impact of changes of government following both national and NT elections, and the absence of funding in the 2014-2015 financial year, also contributed to bringing engagement in Pathways to an inconclusive end.

Systemic racism: Barrier to power sharing and reform
Working across cultures and in partnerships is normally difficult and was inescapable in the Pathways reform project. The difficulty can be attributed to a combination of the underlying power imbalance between Aboriginal communities and government authorities, the participants' separate interests as funders and providers, and the pervasive and perverse impacts of systemic racism. Interpersonal racism was not reported, but systemic racism was considered by many case study participants (both Aboriginal and non-Aboriginal) to be an important underlying influence. Finding good ways to work across race and cultures in this context remains a significant outstanding challenge, in spite of the fact that there was much experience and good will among some of the people involved.
There was a perception that governments did not acknowledge the effectiveness and cultural legitimacy of ACCHOs, or their proper role in shaping the dialogue about community control of the health sector. Although engagement between governments and the ACCHO sector was an essential requirement, there was a mutual perception of failure to maintain commitments and a sense of significant wear and tear on established relationships and mutual trust.
Many of the strengths of the community leaders and representatives, as well as the challenges they faced, were deeply shaped by traditional and contemporary Aboriginal cultures. These aspects, in the wake of the Commonwealth government's Intervention in Aboriginal communities (Evans, 2012), were seen by community stakeholders to have been misunderstood and their significance underestimated.
Risks and other problems that might have been seen as solvable in other contexts became insoluble, resulting in loss of momentum for reform. The perception by some public administrators that governance of Aboriginal organisations is less robust, and that being a champion of reform in this context is a career risk, was evident in the case studies. However, we found no evidence of any formal acknowledgement or discussion of systemic racism or how to counteract it.
We suggest that this difficult situation is made worse by the use of classical contracting by both levels of government. The very contractual instruments that are used to reduce perceived risk undermine the development of robust and responsive PHC programs, give rise to compliance problems, impede the development of strong organisations, and assume a lack of trust (Smith & Smyth, 2010). Change will require not only measures that address the sense of high risk for funders (such as committed political leadership) but also attention to the impact of systemic racism on working relationships. Finding a way out of this conundrum is an outstanding challenge for Australian public administration (Commonwealth of Australia, 2017), although it has been done in both Canada and the United States (Gottlieb, 2013;O'Neil et al., 2016).

CONCLUSION
The need for effective engagement with Aboriginal communities and organisations is well recognised by political leaders and in public policy, and it is acknowledged as being essential for success across all responsible portfolios. The case studies indicate a deep seated incapacity to do so within the current system, given both the unresolved ongoing processes of colonisation and systemic racism, and the limitations of the favoured modalities of contracts and programs. This implementation problem is well known, and we suggest that consistently turning a blind eye to it is itself a manifestation of systemic racism. For public administrators, the task of finding effective, feasible, and acceptable methods for conducting public policy/program development and implementation in ways that address this problem is an outstanding challenge.
Our analytical framework allowed us to develop a deeper explanation of the forces and circumstances that contributed to failure of the intended Pathways reforms, in ways that are informed by the institutional constraints faced by public administrators, and go beyond the usual exhortations to try harder. We suggest that these insights could be developed for use as criteria for designing future reforms in Aboriginal affairs. That is, at the inception when reform projects are being negotiated and designed, proponents should attend to ensuring: 1. That there is recognition of the need for change within government not only at the political level, but also among the senior public servants who are tasked with leading reform efforts for government.
2. That opportunities for the early development of alternative practices are built into the negotiation and planning of reforms, building on any existing niches of endogenous institutional change.
3. That in the process of reform, sustainable or locked-in authorisation is secured, so that adequate time and resources are reliably available, and so that commitment can be sustained for long enough to enable reform goals to be implemented and their impact evaluated. Attention and resources also need to be directed to the development of more permanent alternative institutional forms that enable power sharing and network governance of service systems. This may require different degrees of separation from the business-as-usual of government departments, in the form of a purpose-designed authority or commission.
4. That the challenges of creating effective mechanisms and processes for use by Indigenous organisations/communities and government authorities in working together are acknowledged openly and addressed. In order to counter the impacts of systemic racism and ongoing colonisation, attention is required to power imbalances, to the development and maintenance of trust and to the management of perceptions of risk.