Attributes of effective collaboration: insights from five case studies in Australia and New Zealand

Abstract Developing and implementing social policy aimed at resolving wicked problems has occupied governments and not-for-profit organizations for decades. Such problems are enduring due to their complexity, resulting in a need to harness multiple skills and significant resources to the effort to solve them. Collaboration is one response to this need and is logical because such arrangements have the potential to bring to bear differing experiences, greater resources, and a higher level of understanding. However, collaboration is not easy and, if not done well, can result in significant costs in time and money as well as poor outcomes for those affected. In this paper, we identify the necessary attributes of successful collaborations by examining five case studies that provide important insights into sound collaborative practice.


Introduction
Notwithstanding the primacy of market economics in modern public policy, "collaboration" has become something of a Holy Grail through which multiple parties seek to resolve wicked problems by combining their resources and expertise. Indeed, it is widely accepted that complex socioeconomic problems require experience, knowledge and problem-solving capacities that cannot be provided by any single entity or sector operating on its own (Austin and Seitanidi 2012, 727;Head 2008a;Koppenjan and Koliba 2013, 1;Rittel and Webber 1973).
A decade ago this realization prompted much reflection in Australian public policy circles on options for "networked" approaches to service delivery (APSC 2012;O'Flynn and Wanna 2008). Doubts have been cast in some quarters about claims of a substantive shift from "siloed" to "networked" approaches: some have questioned

The problem
Collaboration often occurs as a response to "sector failure" (Selsky and Parker 2005), thus leading policymakers to focus on hybrid approaches to address complex problems.
Effective partnership-based approaches to public sector governance are based on characteristics such as trust, shared values, implicit standards, collaboration, and consultationas opposed to bureaucratic mechanisms based on control, hierarchy, and chains of command (Favoreu, Carassus, and Maurel 2016, 440). However, collaboration can mean different things to different people (O'Leary 2014, 25-26). In the Australian context, Keast and Brown (2006, 51-52) observe that deliberative endeavors to forge more collaborative ways of working via network-based service delivery models can be jeopardized by failures to make adaptive changes to behaviors, expectations, and processes. Similarly, Shergold (2016, 26) contends that "cross-sector working has been marked by failures of implementation, nerve, and imagination."

Ingredients for success
Fortunately, there is a substantial agreement in the literature about the essential constituents of effective collaboration. As one might imagine, these are presented in myriad ways, but most broadly conform to those identified in Table 1 below.
Above all, a "collaborative mindset" is required (Aagaard 2012; O'Leary 2014) and the appointment of "collaboration champions" can serve an important role in this regard (Crosby and Bryson 2010).
Finally, although public sector partners might represent the largest source of formal authority (power) and resources, they do not need to take on the leading role in collaborative endeavors (Bowden and Ciesielska 2016). Rather they should use their technical capacity to mobilize resources, leverage change, confer legitimacy and ensure sustainabilityin other words, exercise a kind of institutional entrepreneurship and stewardship (Heuer 2011).

The study
This study investigates the dynamics of five multi-party collaborations operating in Australia and New Zealand 1 and aims to identify factors that contribute to outcomes involving public and not-for-profit sector entities. Observations are framed as "findings" from which better practices might be derived.
The primary focus of the research is on the character and quality of relationships between organizations rather than the impact of particular initiatives upon consumers, clients or beneficiaries. However, our underlying assumption is that strong collegial relationships contribute to the realization of policy objectives. It is hoped that this research will generate important, transferable learnings for leaders and managers in the public and not-for-profit sectors.

Research approach and case selection
Similar to previous studies (Keast and Brown 2006), this study involves the review of primary documents pertaining to the establishment and implementation of each collaboration/initiative; semi-structured interviews with persons engaged in the design and implementation of collaborative approaches to the delivery of public benefit; and interviews with relevant stakeholders. They involve institutional and organizational actors with prior structural relationships. (Corwin, Corbin, and Mittelmark 2012;Bryson et al. 2009;Daymond 2015;Jupp 2000;Leat 2009) They are framed around clear objectives and problem definition. (Corwin, Corbin, and Mittelmark 2012;Bryson et al. 2009;Daymond 2015;Jupp 2000, 8;Leat 2009) They have partners with a shared commitment to the objectives having clear understandings of common goals and benefits flowing to partners. (Corwin, Corbin, and Mittelmark 2012;Jupp 2000, 8) They have powerful sponsors providing formal authority; a champion group using informal authority to engage others; and a skilled, committed leadership. (Corwin, Corbin, and Mittelmark 2012;Bryson et al. 2009;Daymond 2015) They have a supportive authorising environment allowing strong relationships with key constituencies through a mix of formal and informal networks. (Bryson et al. 2009;Jupp 2000, 8) They are founded on mutual understanding, respect and trust (accepting that there might be differing levels of trust between collaborators at the outset); and employ explicit strategies for the purpose of building trust and alignment between collaboration partners. (Corwin, Corbin, and Mittelmark 2012;Bryson et al. 2009;Daymond 2015;Jupp 2000;Leat 2009) Their leaders offer catalytic, facilitative and integrated leadership that instils trust and supports the contributions of stakeholders to the collaborative process. Leaders exhibit "bridging" skills (linking to external resources), "mobilizing" skills, "persuasive" skills, and "adaptive" skills (capacity to deal with changing contexts and challenges). (Ansell and Gash 2012, 18;Head 2008b, 739-41;Forrer, Kee, and Boyer 2014, 234;Leat 2009, 28-29;Sloan andOliver 2013, 1860) They have governance mechanisms, processes and structures with final decision making authority that enable the objectives of all parties entering into the collaboration to be known and considered. (Daymond 2015;Bryson et al. 2009) They have a detailed implementation plan including competencies, abilities, technologies or processes required to give effect to the partnership. (Corwin, Corbin, and Mittelmark 2012;Bryson et al. 2009) They have an accountability evaluation system that tracks inputs, processes, and outcomes-including the achievement of common goals and the benefits flowing to each partner. (Jupp 2000, 8;Bryson et al. 2009) They have accepted norms and explicit agreed rules of operation, clear boundaries, structures, and identified roles. (Corwin, Corbin, and Mittelmark 2012;Bryson et al. 2009;Daymond 2015;Leat 2009) They utilize inclusive processes and linking mechanisms to ensure that the objectives of all parties and emerging issues are known, considered and addressed. (Daymond 2015;Bryson et al. 2009) They utilize bespoke processes to reflect the unique and emergent needs of an operating environment characterised by fluidity and hybridity. They are visible and have a public profile; and celebrate and publicise their accomplishments. (Corwin, Corbin, and Mittelmark 2012;Bryson et al. 2009) Despite the prominence of "collaboration" as a major organizing theme in the policy and organizational rhetorics of the public and not-for-profit sectors, and a tendency to refer to all networks as "collaborations" (Mandell, Keast, and Chamberlain 2017, 3-4), clear examples of genuinely collaborative practice proved to be surprisingly hard to find.
The study initially identified twelve initiatives of which eight indicated in-principle willingness to participate (see selection criteria, Table 2 above). Further due diligence led to the exclusion of three of these because: (1) one initiative was not sufficiently advanced to offer significant insight into collaborative practice; (2) one initiative withdrew; and (3) one initiative was excluded when it became clear that its core activities were unlikely to align with the project's research criteria.
The five cases selected for study exhibit considerable diversity. They operate in different jurisdictions, involve different levels of government and operate at different geographical scales. They each have distinct institutional histories spanning a number of policy domains, and are governed with differing degrees of formality. Each involves collaboration between public sector entities, and between public sector entities and external actors, including local community "influencers", civil society organizations or informal community groups. The cases are described in Table 3 below.
Semi-structured interviews were conducted between September 2017 and May 2018. Persons interviewed each played a significant part in the initiation, design and/ or implementation of the collaborative initiative. Interviewees included senior executives, officials, front-line implementers, thought-leaders and the members of backbone/governance groups 2 . In total, 25 individual and group interviews, recorded and transcribed, were conducted with 65 individuals. Written consent was obtained from all participants. Transcripts were coded and analyzed using NVivo software (QSR International, Melbourne, Australia). The data obtained from the interviews was further triangulated with observations drawn from primary documents and from secondary literature (including peer-reviewed and "gray" literature).

What the cases told us
Interestingly, it became clear in the interviews that the collaborative approach taken in each of the cases was not expressly modeled upon an existing collaboration or The study sought to investigate initiatives that: 1. operate across institutional, organizational and/or policy domain boundaries; 2. focus on addressing "wicked" problems and "hard to reach" target groups; 3. reflect a degree of variety in terms of policy domain, scale, size and reach, organizational mission, jurisdiction, and geography; 4. incorporate elements of co-design and/or co-production; 5. exhibit a degree of shared governance between policy leads, commissioning agencies, service providers and service users; 6. have in one way or another challenged the status quo and led to genuine and sustainable innovations in thinking and practice; 7. offer insights on how public sector norms, practices or systems might need to change in order to achieve effective cross-sector collaboration; and 8. demonstrate a set of learnings that might contribute to designing improvements in existing or future projects, even though the case under study might have failed in some way to achieve its objectives. It was determined that cases selected for further investigation should exhibit at least five of the eight characteristics listed above. Overseen by a Working Group Jointly chaired by the CEO of Corrective Services and a nominee from the community sector, Throughcare brokers post-release access to housing, health services, income support, and basic life skills support. The co-chairing model was intended to encourage the community sector to take a primary role as a partner and has had the effect of encouraging strong community sector buy-in as well as helping to make the program more responsive to the diverse needs of its client base. WHO Stops is a community-based initiative that enables local community leaders and members to work together to address complex local drivers of childhood obesity. A major impetus for the partnership came from the work of researchers from the Global Obesity Centre. The initiative is driven by epidemiological evidence of the incidence of, and contributors to childhood obesity and research identifying the costs to the community of obesity.
WHO Stops uses a facilitated community engagement process to: map the causes of childhood obesity; identify intervention opportunities; and convert these understandings into community-built, systemsoriented action plans. WHO Stops acts as an umbrella for subsidiary initiatives in two regional centers in which local backbone groups engage with a wide range of stakeholders to raise awareness about childhood obesity and to stimulate community responses to the problem.
Ten local Children's Teams were established under New Zealand's Children's Action Plan to bring together professionals to assess the needs of vulnerable children using a common assessment approach and, where The Children's Action Plan aimed to address structural and systemic deficiencies identified by a 2015 Expert Panel: service delivery and purchase models that failed to provide a range of effective services and approaches or to be sufficiently A key element of the strategy would be "a single multi-agency plan for each vulnerable child, with someone taking responsibility for overseeing it and ensuring it is implemented". Oversight for each Children's Team is provided by a Local Governance Group consisting of senior managers from core government agencies (continued) upon an established model of collaborative practice. Perhaps predictably, persons interviewed exhibited little awareness of the academic literature concerning collaboration although many had some knowledge of Kania and Kramer's (2011) work on collective impact. Even so, the five cases investigated for this study offer findings that are broadly consistent with the characteristics of successful collaborations discussed in Table 1.
The cases also exhibit a number of important features that are not extensively canvassed in the previous literature. These additional factors for collaborative success are addressed in the sub-sections below.

Each collaboration should be specifically designed to address the specific problem
Each of the initiatives we examined is seeking to address problems that have proved resistant to conventional policy approaches. Thus, it seems clear that recourse to a collaborative approach is a response to the twin effects of complexity and sector failure. However, the collaborative solution must be designed specifically in response to the presenting problem.
A major strength of collaborative approaches is the capacity to allow for the crafting of fit-for-purpose local solutions addressing local priorities whilst allowing for local stewardship. Whereas bureaucracies tend to favor more uniform approaches that are amenable to consistent and comparable impact metrics, collaboration requires an acceptance of complexity, uncertainty, and diversity of approach.
Consequently, it is not feasible to impose a standardized collaboration framework with little regard to local circumstances and local aspirations. As one interviewee remarked on an attempt to impose collaboration from the top: It was just a blanket approach. "It's all about collaboration. This is the focus of this rollout and everyone gets an even split of the pie," as opposed to identifying what the issue is, identifying where you can make the biggest difference and then factoring that in those areas and regions.

Senior managers must support collaboration by allowing it to work
A collaborative process needs to be formally supported and authorizedand, importantly, understoodat the executive levels of partner organizations. This usually involves the establishment of what is referred to in the collective impact literature as child-centred; a siloed system with insufficient partnership and collaboration around children's needs; restrictive funding approaches that do not permit innovation or the creation of sustainable services to meet changing needs; and diffuse accountabilities across various agencies. and, where appropriate, other key partners such as non-government organizations, Iwi and local government representatives. It is hoped that active collaboration will become "business as usual", replacing modes of working based on parallel working and ad hoc communication.
a backbone organizationor its functional equivalentto support the administration of the collaboration and allocate resources (Kania and Kramer 2011).
Critically, the collaboration also needs to be based in an operating environment that allows for flexibility, experimentation, distributed governance and decision making, and control over resource allocation and deployment. If the bureaucratic systems of partner organizations are permitted to exercise traditional gatekeeping roles with respect to decision-making and governance arrangements, the collaborative process can easily be smothered, leading to collaboration failure. Indeed, Kotter (2012, 2-6) points out that while traditional hierarchies and managerial processes do many things well, they are not sufficiently nimble to "address the challenges produced by mounting complexity and rapid change." The solution is a "dual operating system" comprised of a "management-driven hierarchy" and a second operating system or "second operating space", "devoted to the design and implementation of strategy, that uses an agile, network like [sic] structure and a very different set of processes" connected to the hierarchy via people who populate both spaces, liberating information from "silos and hierarchical layers" (Kotter 2012, 2-6).
Formal authorization for collaboration occurs in the "primary operating space"where the dominant, normative operating culture of partner organizations resides, whereas collaboration itself occurs within the "secondary operating space". It is essential that this second operating space is "treated as a legitimate part of the organization" and does not come to be viewed as a "rogue operation" or "the hierarchy will crush it" (Kotter 2012, 8).
Our observation is that collaboration appears to occur within curated spaces very much like Kotter's "secondary operating space." Different rules, values, and behaviors apply in secondary operating spaces, where fidelity to collaborative aims takes precedence over the formal mandates of partner organizations. As expressed by one interviewee: [Honesty and trust] is really important … You've got to feel safe and you don't judge. We might not agree, but you don't judge as a result of that. We just don't agree. That's fine … It tests our thinking. It enables reflection, it's all that sort of stuff. So that's about building a team that's united and unified most of the time.
It is important, however, to agree and document decision-making processes amongst collaboratorsa flexible secondary operating space does not mean an operating environment devoid of administrative discipline.

Develop policy and operational tools that work for collaboration
We have already observed that collaboration is often identified as an important ingredient when addressing wicked problems. However, it is difficult to implement in practice due to accepted norms of governance and management. Indeed, Canadian social entrepreneur, Mark Cabaj, contends that collaboration is not counter-intuitive: rather, it is counter-cultural 3 . The key challenge going forward, according to Cabaj, is to cultivate a set of policy and operational tools that will support the establishment of secondary operating spaces. One interviewee told us that, ideally, collaboration is "outward facing" and focussed on mediating and accommodating a range of perspectives and priorities, as opposed to being inward-facing and narrowly focused on fidelity to process. This is uncertain and challenging organizational terrain for those charged with the task of collaboration.
Collaboration often exists in dynamic tension with the dominant culture of partner organizations. This is both a stimulus for innovation and problematic for organizational logic. As one interviewee observed: it gets confusing within the community and within [partner organisation's] own staff about who "owns" this … And the Shire and the [another partner organisation] at times have been a bit quietly annoyed about "is this a collaborative approach or is this all about [partner organisation]". So there are those kinds of ongoing tensions that we feel like we take two steps forward and one step back.

Understand the relationships between, and culture of, collaborators
A common theme emerging from the interviews concerns the importance of a learned set of skills that are vital in forging and sustaining productive collaborative relationships. Collectively, these might be called "collaborative intelligence" (CQ) which combines an acute sensitivity to the interpersonal dimension of human interaction with an astute appreciation of the emotional resonance of systems, relationships between organizations, and the "baggage" brought to the collaboration by various actors (Butcher 2017).
In a very real sense, most of the people interviewed for this study have learned the fundamentals of collaborative practice "on the job". Although many people working in the public and community sectors exhibit a capacity for high CQ, differences in organizational culture can encourage or inhibit its expression: I think collaboration is often misunderstood as something that you just do. Occasionally I hear the phrase, "barriers to collaboration", as if you've just got to break the dam and it will flow naturally. But, in my view, collaboration is a learned set of skills. It's hard; it's complex; it happens at various levels in various ways. It can happen a bit or it can happen in a very deep and enduring way. I think having a group of people that, if you like, learn on the job together how to collaborate was really critical to the success of this both in development and implementation.
As for the attributes required of "intelligent" collaborators, one interviewee observed: I think you need to have people in those roles [who] understand the ecosystem and how the bits work and how those levers work, some of which are informal levers and some of which are formal levers. But they also have to be able to actually understand from a transactional level how to get things done and to make sure that happens.

Collaborators must communicate with bureaucracies to give assurance
It is not enough for authority to cascade from the top. Sustaining collaboration requires reciprocal flows of assurance: between partners; between partners and stakeholders; and between partner organizations and frontline workers: For some of the bureaucracies it is that "We want the community to help us do what we need to do"and there's nothing wrong with thatbut "We don't want them to play the game any other way except the way we play it." Effective community work means that you take a stick to bureaucracies and you get them out of the way and realize they're a resource and not the main game. 6.6. Middle managers within bureaucracies needs to be recruited to the cause One clear message from those engaged in collaboration is that they frequently meet middle management resistance and that such resistance often occurs despite clear directives from executive management giving sanction to collaboration. It is at this level, perhaps, where the dominant incentive structures reward territoriality, conservatism, risk aversion and excessive focus on outputsall qualities that militate against genuine collaboration. It cannot be assumed that support for collaborative approaches exists at all levels within partner organizations. Because communication strategies around collaboration tend to focus on external audiences, it is possible that internal messaging tends to be neglected.

Have realistic time frames in mind
The collaborations examined all exhibited long lead times for design and implementation, involving intensive and complex processes of relationship building, establishing legitimacy and trust, collectively framing the problem, and agreeing on ways of working. Cutting corners on the establishment and relationship building process will build significant problems into the system. One interviewee told us: I think one of the key messages that's been good for people is that it is a four to fiveyear journey.
Another interviewee remarked of the process leading to the establishment of a local multidisciplinary team: we realised, "Gosh, just to set this up took way longer than we ever imagined," just getting people on board with the concept of it let alone to actually come together and work together and actually achieve some results.
6.8. Consider how the collaboration will demonstrate impact early but realistically A recurring theme in each of the cases is the importanceand problematic natureof "evidence of impact." Interviewees generally agree that it is very difficult to demonstrate impact in the early stages of collaboration. This is in large part due to the significant upfront investment of effort in building relationships with partners and other stakeholders.
It is generally acknowledged that the design of social programs should have a sound evidential base in order to persuade decision-makers, authorizers and stakeholders about the soundness of the approach. In the early stages of collaboration, partners sometimes struggle to reassure authorizers that they are making progress because evidence that relationships are being built is not accorded the same value as conventional measures of impact.
Authorizers often underestimate the investment of time, effort, and emotional energy required to create viable, constructive, respectful relationships between stakeholders and become impatient to see measurable demonstrations of impact. Authorizers often do not consider that the creation of a new service delivery culture or the establishment of new forms of collaborative relationships offer evidence of impact. As suggested by one interviewee: I get annoyed sometimes about, "Can you evaluate it? Can you tell us what's happening?" These things take time in terms of how you manage them. And it takes away the human context. Also, because collaboration often involves charting unknown territory, established measures of effective collaboration are not well-understood by frontline practitioners who might have poor access to the sizable and diverse academic and practice literature.
6.9. Consider how the collaboration will be sustained over a long period Collaboration can be difficult to sustain. Indeed, its informal or semiformal nature suggests impermanence. Also, formalizing or normalizing collaboration can undermine its dynamism and sense of collective purpose, while burnout and high staff turnover contribute to the erosion of corporate memory.
It can be difficult to maintain the founding purpose, sense of mission and personal commitment that saw the collaboration get off the ground in the first place. Any dilution of the capacity for collaborators to share and find validation in their foundation stories can diminish their collective sense of mission and commitment to change.

Conclusion
Interest in collaboration reflects a growing recognition that the traditional bureaucratic model of public administration is not up to the task of addressing complex social problems. Collaboration most often occurs in dynamic environments in which multiple logics are in play. And, although effective collaborations have a number of things in common, all collaborations are in some respects unique. This is because a collaborative endeavor is not primarily an entity or an organizational structureit is a set of organizational and interpersonal relationships shaped by the nature of the problems being addressed, the predispositions and capabilities of key actors, and the characteristics of the places in which the problems occur.
A number of important themes have emerged from the interviews undertaken for this study. First, collaboration as a response to wicked problems needs time and dedicated resourcing. Second, the trajectory for collaboration can be unpredictable and requires a capacity to tolerate a lack of certainty. Third, one cannot underestimate the time, effort and emotional energy required to manage internal and external relationships; maintain the internal integrity of the process; and ensure the external legitimacy of the collaboration. Fourth, personal dedication and commitment to the issues at hand is critical for maintaining focus and effective collegial relationships. It is also what sustains participants in the process when the going gets tough. Fifth, partners not only need to instill, and sustain, confidence and goodwill, they need to provide appropriate assurance to their executive and board (and upwardly support the executive and board who might themselves be called upon to provide assurances). Collaboration also needs to be outward-looking and able to offer assurance to a range of external stakeholderssome of whom might have perspectives that are not fully aligned with the organizing themes of the collaboration. And sixth, collaboration has an organic quality; goalposts will change, thus requiring a capacity for nimbleness and adaptability. For this reason, formal terms of reference are useful as starting points, but might unduly fetter collaboration practice.
We consider that the foregoing observations are equally applicable to collaboration regardless of the sector in which it occurs. Although each of the cases is sited substantially (but not exclusively) in the social sector we contend that the fundamentals of effective collaboration are essentially "sector agnostic." Finally, it is important to bear in mind that collaboration might not always be the most appropriate strategy for every problem. Considering the challenges involved, and the risks of collaboration failure, decision-makers would be wise to not look upon collaboration as a talisman capable of shepherding them safely through complex policy terrains.