Reflections from the Think Tank Initiative and their relevance for Canada

ABSTRACT From 2009–2019, the International Development Research Centre (IDRC) managed the Think Tank Initiative (TTI) as a contribution to strengthening applied research systems in Latin America, East and West Africa and South Asia. TTI’s experience supporting these organisations yielded interesting insights into how collaborations between researchers and civil society organisations can be made more effective. If we wish to see greater benefits from researcher-civil society collaborations in Canada in support of international development policy and programming, it might be time again to reflect on how to address the gap created after the closure of the North–South Institute.


Introduction
The idea that collaboration between academics and civil society is a good thing assumes that well-functioning systems of research and knowledge sharing can improve development policy and practice, which should in turn lead to better development outcomes. This view reflects the distinctions that were made long ago by Aristotle on the nature of different kinds of knowledge. Theoretical knowledge aims to establish truth. Poetical knowledge helps produce beauty. Practical knowledge (praxis) is a guide to action. Nineteenth and twentieth century philosophers built on this idea to form notions of how philosophy could contribute to social change, and today, these ideas are deeply embedded in intellectual frameworks behind philosophies across the fields of education, health and even some spiritual practices. Thinking about ways of generating praxis and the reflective processes and engagements with the "doing" that it requires, can help think through not only the value of certain collaborative efforts, but also how they happen and how they can be sustained.
The Think Tank Initiative (TTI) encapsulated one way of thinking about this kind of collaboration. The TTI Theory of Change (ToC) was derived from the view that independent policy research institutionsor think tanksare in a unique position to effect positive change in their societies. By generating and analysing credible local data, they can enhance public policy debates and promote more objective, evidence-based decision-making that makes real, sustained improvements in people's lives. (Think Thank Initiative 2020) If research-based knowledge, evidence and data is to improve development policies and practices (and contribute to better and more lasting development outcomes), then it is important to strengthen key actors who play a vital role in making this happen. In the case of TTI, the programme was designed to provide think tanks with the means to strengthen themselves and set their own policy-focused research agendas. The assumptions behind TTI's theory of change held that strong think tanks, with missions and mandates to generate praxis, are uniquely positioned in knowledge systems, able to bridge the gap between research and action in a timely and effective way. They perform similar functions to universities and NGOs but their delivery is different, and they are often able to take risks that others do not (Foresti and Hedger 2015;de Boer 2015). They thus perform a useful and catalytic function in knowledge systems (Weaver and McGann 2017). Public policymaking can improve when there are think tanks who are "grounded in the importance of evidence, but are entrepreneurial, responsive and politically savvy in the use of that evidence" (Lodge and Paxton 2017).
Insights from TTI and the organisations it sought to help strengthen should serve as interesting points of reference for those interested in furthering collaboration between researchers and practitioners in the Canadian international development community for two reasons. First, observing first-hand how the think tanks supported by TTI were able to manoeuvre in their contexts and contribute data, evidence and analysis into national economic and social development processes suggests that think tanks as an organisational form can be effective models of applied knowledge-driven social collaboration. What is interesting is not what these think tanks focused on but how they did it, what this suggests about conceptions of collaboration at the nexus of research and action, and why there are persistent efforts to fill the "gap" left by the demise of the North-South Institute. Second, lessons from TTI's efforts to help strengthen these organisations can help those who might seek to fill the gap in the Canadian context by providing some ideas about what the nature of the current gap is, and what kind of an organisation might effectively fill it. They may also help in thinking through ways of addressing some of the obstacles that exist to sustaining effective, independent policy research organisations with an international development focus in Canada.

Background
The Think Tank Initiative was a ten-year, multi-donor funded programme of organisational strengthening that supported 43 independent policy research organisations in 20 countries in Latin America, West and East Africa and South Asia (see Table 1). The International Development Research Centre was both a co-funder and implementer of this initiative. The support provided by the Initiative consisted of a package that included flexible, long-term (10 years) grants, accompaniment by regionally based programme officers, additional training and technical assistance, and opportunities to learn from and collaborate with peer organisations across the Global South.
The organisations TTI supported were those whose missions and mandates focused on informing and influencing development policy and practice. What united them was their independence (relative to their contexts), and their focus on informing policy or practice through research (or the desire to increase their focus on research). Beyond this, there was a great deal of diversity in the cohort, and organisations came in many shapes and sizes. Most focused exclusively on public policy in their own countries. 1 The "think tank" moniker was not necessarily an apt term since for many it came with American associations or did not fit self-perceptions (Datta 2011). The organisations activities stretched from more academic research and policy discourse framing to knowledge brokering, stakeholder convening and collaborating, advocacy, and policy training. Some like Grupo de Análisis para el Desarrollo (GRADE) or Instituto de Estudios Peruanos (IEP) in Peru or the Kenya Institute for Public Policy Research and Analysis (KIPPRA) in Kenya and REPOA in Tanzania had existed for decades when they were selected to join the program in 2008/2009. Others were newer organisations such as the BRAC Institute of Governance and Development (BIGD) in Bangladesh, which resulted from a merger between two pre-existing organisations, or were relatively small and new, such as the Science, Technology and Innovation Policy Research Organization (STIPRO) in Tanzania or the Center for Study of Science, Technology and Policy (CSTEP) in India. Several were university-based research centres or had close organisational relationships with universities (the Institute of Statistical, Social and Economic Research, ISSER at the University of Ghana; Makerere Institute of Social Research, MISR and the Economic Policy Research Centre at Makerere University in Uganda; BIGD with BRAC University in Bangladesh) while others were the research outfits of professional associations (the Institute of Economic Affairs, IEA in Kenya and the Ethiopian Economics Association, EEA in Ethiopia). One was even situated within the executive branch of government, as was the case with the Ethiopian Development Research Institute (EDRI) in Ethiopia, which sat within the Prime Minister's Office. 2 This diversity of form but relative commonality of purpose was reflective of the broader community of independent policy research organisations in the Global South, and indeed around the world.
These organisations also occupied different social and political positions within their countries, regions and globally. Some like EDRI were extremely close to centres of power, while others such as the Indian Institute of Dalit Studies (IIDS), by virtue of their focus, were less so. 3 Still others faced challenging political contexts for some or all of their time in TTI, such as INESAD and Fundacion ARU in Bolivia or Grupo FARO in Ecuador. As a result, each of their approaches to engaging in development processes in their respective countries varied considerably.
While not all can be considered civil society organisations (CSOs), a significant number of TTI think tanks either self-identified as CSOs, had important alliances with CSOs, or as a result of their beginnings, saw connections with citizens, communities and CSOs as central to the way they worked. These included: . Centre for Budget and Governance Accountability (CBGA) in India, who have done pioneering work on citizen engagement in budget monitoring. . Public Affairs Centre (PAC), based in Bangalore, India, whose reputation was built through the development of social accountability tools. . Institute for Social and Environmental Transition (ISET) in Nepal, who sees its mission as helping to connect communities with policy makers, and to articulate citizen demands and interests in ways that marry up with science.
. The Centre for Poverty Analysis (CEPA) in Sri Lanka, whose roots lay in advocating for marginalised communities on issues of human rights, housing and gender equality. . Advocates Coalition for Development and Environment (ACODE) in Uganda, that started life as an organisation of lawyers advocating on behalf of communities. . L'initiative prospective agricole et rurale (IPAR) in Senegal, a younger sibling to the more established Consortium pour le Recherche Économique et Sociale (CRES), whose route to research work came through civil society (whereas CRES' came more through academics). . The Centre for Population and Environmental Development (CPED) in Benin City, Nigeria, with strong commitments to addressing pressing social issues through its research and engagement with both communities and state governments. This is not to say that the other organisations in TTI had no engagement with civil society. Many that did not self-identify in this way, or who did not systematically engage with groups other than policymakers, still broadened their engagement efforts beyond the state sphere. Others developed new research agendas, which necessitated reconstituting connections with different segments of society.
The nature of the changes and journeys undertaken by the think tanks supported by TTI was captured in TTI's external evaluation final report (Niras Development Consulting 2019). The evaluators pointed out that organisations TTI supported began at different points, reflecting their context and the unique histories of their organisations. Some had existed for decades and had significant numbers of researchers (over 50) while others were like start-ups, consisting of only a handful of staff. Their relationship with governments and other policy actors varied tremendously as well, as did their previous experiences with unrestricted funding of the kind TTI provided. Nevertheless, the evaluators discerned certain patterns of change within the TTI cohort, of which TTI support was but one of many contributing factors. Some organisations went through transformational change, reflected in how they were able to diversity their sources of revenue, the number and quality of their staff and their ability to produce research and engage with key policy constituents. Others went through accelerated change, typically starting from an established base but taking advantage of TTI support to invest in areas like communications that allowed them to become more effective. A third group were able to consolidate their positions within the evidence-to-policy ecosystems, becoming more financially secure in the process. A final group were only able to survive in the face of difficult circumstances, typically resulting from governments that were hostile to civil society organisations and/ or to policy advice emanating from outside the state.
Amongst these categories, there was no single, successful model of stakeholder engagement, and even within the same country think tanks took different approaches to informing and influencing the same policy issues. This suggests that there is no single model to follow. The context can shape whether organisations focus on a narrow group of constituents within the state (an "insider" approach) or collaborate with a wider set of stakeholders beyond (an "outsider" approach) (Lodge and Paxton 2017). Creativity and flexibility in making different kinds of connections characterised 296 A. W. HURST many of the organisations who were effective in achieving their chosen ends in their own contexts. For many, it appears that the continuity an organisational locus provides for these kinds of connections was useful in sustaining them and making them effective in shaping policies and informing practices.

Collaboration between researchers and civil society amongst TTIsupported organisations
Interpreting the journeys the organisations underwent during their time in the TTI cohort, and how this related to their local contexts, yields some interesting insights into how think tanks can help foster collaboration between those academics concerned with praxis and civil society organisations concerned with finding ways to scale up the impact of their efforts (whether through policy influence or programming). First, the nature of a think tank's position within or engagements with other social and political actors (whether elements of the state or civil society), was heavily influenced by the broader political economy of their domestic context, and by the nature of the policy issues they chose to work on. In most TTI countries, governments and states were strong, or aspired to be. This reflected in part the prevailing preference for "development state" models of governance, and political projects by governments that viewed civil society actors with suspicion. In countries like Bolivia, Ecuador, Ethiopia, Rwanda, Tanzania and Nepal, space for organised, non-state actors was limited, or became limited, for significant periods of time. This in turn required adjustments in how think tanks approached their work. In such instances, think tanks had to navigate complex relationships with the state. Even those organisations with the deepest roots in civil society typically had to expand their engagement and outreach efforts to other parts of society, for instance with business, unions or professional associations. In some cases, this was to demonstrate their relevance to state actors and represented an unavoidable pathway to influence. In other cases, where governments were openly hostile, broadening engagements with actors beyond the state reflected efforts to find alternative, less direct pathways to influence. In all cases, their engagement efforts were constantly shifting, and represented a delicate balance between the principles that guided each think tank in their actions, and the need for pragmatism if they were to have influence.
At the same time, the accepted wisdom around certain policy issues, and the associated space for civic engagement on them, shaped these engagements and collaborations. Research on issues considered "technocratic", such as macroeconomic or industrial policy, were often considered a topic for experts and government policymakers, and not amenable to research that embraced a more broadly deliberative or participatory approach. That said, some of the most interesting work was done by organisations challenging this approach. For example, CBGA in India sought to provide the analytical basis for civil society organisations to engage in budgetary processes at the state level in India. In contrast, research on other policy issues, such as health or education, where there has long been an expectation of citizen engagement and participation, saw think tanks naturally seek out or involve participation from a broader spectrum of society.
Second, the positions of influence these organisations were able to achieve, and the impacts they were able to have, reflected the nature of the research systems in their countries, and research orientations arising from their missions and mandates. Most of the think tanks supported were in countries where universities' research capacity was weak to non-existent. The reasons for this are many and varied, but it meant that these organisations were not displacing or competing with other national research institutions. In fact, they often contributed to their improvement, becoming essential parts of the broader ecosystem of research and knowledge in the countries in which they operated. 4 The primary role of universities in almost all of these contexts was teaching, not research. This has left a gap that many think tanks have been able to fill.
More importantly, the missions and mandates of many of these think tanks meant their research was more applied, and often less academic in nature. Many think tank Executive Directors (EDs) noted that their researchers often preferred think tanks to universities, and not necessarily because the pay was any better (it was often more work and more insecure). Rather, many researchers wanted to "make a difference" and "contribute" to the betterment of their countries and societies, which they felt was difficult to do from universities or by publishing exclusively in academic journals. Many relished the opportunity to engage in policy processes through their work, seeing this as likely to have greater impact than university-based research where institutional norms around value, such as publishing in peer-reviewed journals, have little to no bearing on how important or relevant to pressing development problems the research is. By contrast, many of the think tanks possessed, either implicitly or explicitly, a broader view of research quality that reflected the importance of the research topic relative to local problems, the legitimacy of the research process that produced the knowledge, and the positioning of research for use by policymakers and practitioners. 5 On the latter two points, from the perspective of many (not all) think tanks, an important dimension to ensuring legitimacy and positioning for use in the research process, was making connections to, and engaging with, citizens and communities, whether through the research process itself or through advocacy efforts with other CSOs. While TTI as a programme, as well as some of its partners in Africa, engaged in wider discussions on such issues (Hayter and Taylor 2018), TTI did not set out to foster such connections in any structured way. Nevertheless, it is likely that such engagements, and the ability of organisations to craft incentive systems that encourage them, are easier to undertake from within the flexible organisational structures of think tanks, compared with the heavier and more bureaucratic institutional structures of universities. In a sense, these different engagements provide evidence of the socially and politically connective opportunities think tanks offered, generating collaborative possibilities that "worked" within the constraints of their contexts and constituting spaces for researchers and practitioners to work towards common ends.
Third, collaborations at the nexus of research and actionwhere researchers and other social actors meettakes time and money (Hurst 2019). This is an obvious point, but its importance bears restating. Whether it is full co-production of knowledge (Oliver, Kothari, and Mays 2019) or other "lighter" forms of citizen engagement (Hayter and Taylor 2018), collaboration has costs that need to be borne. TTI's core funding provided these organisations with the flexibility and independence to set their own agendas, and ensured they were able to cover the costs associated with public engagement and outreach in all its formsfrom dedicated communications and outreach units staffed by individuals trained in these areas, to face-to-face dialogues that brought together different stakeholders around issues of common concern. Prior to TTI, many think tanks were not good at apportioning and charging these costs across their research portfolio and were too reliant on too 298 A. W. HURST few funders. TTI's support helped many organisations unpack the costs of producing research to inform action and to improve their costing of these practices. While getting a handle on these costs was important and getting organisations to recognise that they represent a key dimension of the research process was not always easy, now that many of the think tanks that TTI supported are convinced, they face the challenge of figuring out ways to cover these costs going forward, given that TTI funding has ended. TTI's core grant gave organisations the flexibility to pay for some of these mission-critical core functions; in contrast, many other funders will not pay for these costs. In some cases, there are institutional obstacles that cap indirect costs of doing research. In other cases, there is an unwillingness to accept such costs as a necessary contribution to achieving the influence and impacts that donors expect from the research projects they fund.
Across all contexts where TTI worked, there was (and still is) a clear power imbalance between funders and those they fund. This imbalance can never be fully overcome; but the narratives that shape funding decisions can be challenged and modified. Everybody acknowledges that the "results agenda" is necessary, especially for bilateral funders who are accountable to tax payers and in the context of a growing donor focus on demonstrating value for money. However, most also acknowledge that this has distorted the development narrative and privileged simplified, technocratic views of "how change happens". This has made it difficult to defend messier, less linear change processes at the heart of development, and therefore harder to demonstrate the relevance of such costs to results. In some cases, TTI was able to help some think tanks develop better costing models and generate better arguments to defend these costs with funders, which in some cases helped organisations renegotiate their contribution agreements. This remains a big challenge however, and researchers and civil society actors alike must help shape new narratives that can open up discussions of what constitutes effective contributions to development processes, what costs are reasonable, who should pay for them, and how the effectiveness of such contributions should be measured (Niras 2019). 6 In short, TTI showed that think tanks can be effective and efficient ways to overcome the challenges at the research/practitioner nexus. They offer the potential for researchers to be more effective agents of change through the collaborative opportunities they open up, allowing linkages with those actors (government, civil society, media, private sector) that can make change happen. The assumption that local research organisations embedded and engaged in their national contexts are best able to navigate and negotiate opportunities for policy influence remains compelling when considering the experiences of TTI-supported think tanks. These experiences suggest that think tanks can facilitate evidence-informed policy change and action by virtue of creating dedicated collaborative spaces and leading deliberately constituted processes where researchers and other relevant social actors can come together to talk about the nature of development challenges and work through, in a structured way, possible ways to address them.

Reflecting on insights from TTI and their relevance for the Canadian international development context
Since the North-South Institute (NSI) stopped operations in 2014, there have been many discussions about what contributions Canadian research and researchers should make within Canada and beyond, and how best these contributions can be made. A gathering of interested representatives from Canadian government, universities, international think tanks and civil society organisations in 2017 agreed that there was a "gap" in the Canadian development context. There was insufficient, dedicated national capacity to engage with international counterparts, help Canadians understand and interpret global development issues and processes, and in turn provide Canadian perspectives on these same issues and processes. Participants felt this gap should be filled by some kind of organisation or network, although with exactly what mandate and in what form there was no consensus (Martel 2017).
The Canadian context is very different today than when NSI was founded in 1976, different even than 2014. The arrival of the internet has dramatically expanded access to research and knowledge for Canadian international development policymakers and other development stakeholders. With the click of a mouse, Canadians now access to top quality development research from across OECD countries and increasingly, from the Global South. The creation of new, development focused think tanks like the Center for Global Development (2001) and the expansion of existing ones like the Overseas Development Institute has generated more research competition, and mirrored the globalisation that has served as the context for their research. More recently, the advent of social media has created new ways for non-Canadian development research organisations and researchers to occupy the space once filled by NSI. Those civil society organisations with the interest and means of investing in research have been multinational in character, with research often centralised or largely generated in the non-Canadian organisations, as was the case with Save the Children and Oxfam.
At the same time, a decade of Conservative governments (2006-2015) who did not see the value in supporting international development research or research organisations in Canada led to significant funding cuts for the sector. Organisations like NSI and Rights and Democracy appeared to be too dependent on government funding and unable to pivot in the face of new political realities. A small philanthropic sector in Canada insufficiently interested in international development did not step in when the government withdrew funding, Jim Balsillie's support for the creation of the Centre for International Governance notwithstanding, Rightly or wrongly, in the post-NSI context, Canadian development research has come to be seen on the whole as either too academic, too predictable, 7 or not relevant to the Canadian development policy questions of the day. The organisations that have survived, such as the Canadian International Development Platform, are making valiant efforts but of limited scope. As shown in the Introduction of this special issue, 8 some civil society organisations in Canada continue to conduct some research in support of their advocacy (Oxfam Canada, the Canadian Council for International Co-operation-CCIC) but they cannot match the substantial research departments of their international counterparts, which are staffed with highly qualified experts who are well connected to interested and research-attuned programme colleagues and, until recently, were well funded. Some eminent and long-dedicated independent researchers who contribute to efforts to influence policy through force of will, reputation and specialised research-based advocacy efforts (e.g. The McLeod Group) persist but they lack platforms that can sustain the concerted levels of engagement required to be effective in the current context. There are real limitations when research collaborations happen in an ad hoc, piecemeal and timebound fashion. The fact that Canada's ODA to GNI ratio continues to shrink is perhaps the starkest evidence of these limitations.
Against this history and present, TTI's experience is relevant to the Canadian community of international development researchers and actors in several ways. The first relates to ways of thinking about what constitutes research excellence in development focused research in Canada. 9 Drawing from years of experience funding research in developing countries, IDRC has developed an approach for understanding and evaluating quality research for development that reflects the critical role of local context and recognises the importance of going beyond factors like methodological appropriateness in understanding the value of applied research . These other critical dimensions include research legitimacy (such as gender responsiveness, inclusiveness, engagement with local knowledge), research importance (originality, relevance) and how well knowledge is positioned for use (knowledge accessibility and sharing, and its timeliness and "actionability"). In the context of collaborative efforts, researchers in Canada need to reflect on their researchwho it is for, how it is conceived of and undertaken, how and when it is communicated and on what measures its quality should be judged. Publishing in a "high impact" journal available behind a paywall or publishing a 200-page peer reviewed book is likely not a relevant measure of research quality in this context. The incentives and measures of performance for university academics are ill suited to generating praxis and cannot adequately enable research that contributes to the Canadian international development context, or for that matter, to development thinking and practice outside of Canada.
Second, Canadian development research needs (a) context appropriate vehicle(s) that can bring together development stakeholders in government, civil society and the private sector in dynamic and innovative ways. The experiences of TTI suggest that fit for purpose independent policy research organisations (aka think tanks) can meet this need by creating spaces to find common ground where possible, establishing research agendas that meet the real needs of policymakers or practitioners. These organisations can mobilise the resources to make this happen and help communicate how their efforts can make a difference. In the past, the geographically dispersed nature of the Canadian development community was an obstacle to establishing a critical mass of scholars and practitioners, such as exists in the south of the UK. Today, this should be less of a problem given new communication technologies and the orientation of Canadian policy makers to (in theory) look beyond Ottawa for inspiration and ideas.
To be sure, there are other models besides think tanks, of which several have emerged over the last decade (the Centre for International Policy Studies at the University of Ottawa, the Munk Centre at the University of Toronto, and the Asian Studies Centre at the University of British Columbia, are all good examples and are challenging the Norman Paterson School of International Affairs' traditional monopoly and role in this area). However, they do not have clear mandates or international development research priorities and their connections tend to be with senior policy makers. This can be a good influence strategy but provides little access for broader stakeholders like civil society actors and thus the quality of their research (as defined above) can be questioned. Canadian NGOs could look to establish more robust research units (as the big UK NGOs like Oxfam and Save the Children have done), either individually or as a group. The organisational form will follow the functions, which in the current context come together in ways that were not necessary or not possible at the height of the North-South Institute's influence. Big organisations like CGD or ODI could open up "branch plants", bringing their considerable names and creating local platforms that might better serve the Canadian context than their current sporadic or virtual engagements offer.
The fact that many of these existing attempts to fill the Canadian development applied research gap are in universities speaks to the challenges of financing such organisations. Funding is unquestionably a challenge and for certain a difficult one. There is no single institutional home within the federal government system to foster collaboration between the international development research community and civil society. The Next Generation programme notwithstanding, IDRC does not typically fund this kind of work, and as Next Generation research has highlighted (Martel, King, and Baruah, forthcoming), research granting councils are not well set up to do this eitherat least with a focus on the international. Canada's Policy for Civil Society Partnerships for International Assistance is a promising opportunity, if resourced, to connect a research to action dimension to any implementation plans that are developed for the policy. But financing should not be the starting point when looking to address the gap. The first step is making a credible and compelling case for why such organisations are needed, and what value they can bring, either to furthering Canadian development praxis or contributing Canadian perspectives to global development discussions and efforts. The experience of many of the think tanks supported through TTI suggests that research-to-policy-and-practice systemsin effect, systems whose actors have the capacity to learn and adaptwork better when catalytic actors that can broker applied knowledge exist and are able to animate such processes in committed ways for sustained periods of time.
To be clear, what is not needed is an organisation of Canadian researchers whose focus is on conducting research in developing countries. There are plenty of researchers in the Global South doing this already, and the days of policy expertise residing in OECD countries of relevance to countries in the Global South are waning. A new organisational form (call it a think tank or call it something else) could easily connect with existing networks of researchers and think tanks in the Global South, drawing upon genuine and grounded research happening there. However, there is arguably still a need for research in Canada on Canadian development policy and practice, done by researchers who understand the Canadian context as it relates to national and global development processes. Unlike "last generation" think tanks like NSI, a new Canadian organisation (or organisations) would thus have a brokerage function alongside its Canadian focused research efforts.
History should warn against an excessive dependence on government funding for this kind of effort. This makes it challenging for Canadian development think tanks, since the Canadian philanthropic sectorwith a few important exceptionsdoes not yet provide funding opportunities that would provide a counterweight to Canadian government funding. But these obstacles are faced by researchers and civil society organisations in every country. The best collaborations are those that can figure out how to position themselves for influence under such constrained circumstances and articulate principles of research excellence that underscore why what they do matters for the Canadian development community and the global community of development practitioners.

Conclusion
Think tanks are not a panacea, and certainly not the only solution to addressing complex problems through knowledge mobilisation. But they can be helpful agents 302 A. W. HURST of social change and there are clear ways they can add value in the current challenging context (Mendizabal 2017;Hernando, Stone, and Pautz 2017), producing research whose quality goes beyond traditional academic measures of rigour. Contributing to social change is presumably the overarching reason there is such interest in collaboration between researchers and civil society. In the Canadian context, think tanks may be one way to help improve such collaboration. There is a clear and acknowledged gap in the Canadian think tank community (Biggs et al. 2015). The experience of the think tanks involved in the Think Tank Initiative holds interesting lessons in thinking about how this gap might be filled. This experience also suggests the sustainability of an effective organisational arrangement, whatever form it might take, will depend on it maintaining relevance and legitimacy in the Canadian context first and foremost.