1 THE RISE OF CITY–UNIVERSITY PARTNERSHIPS

There was a time when fierce town vs. gown battles were fueled by class differences and cultural conflicts (Warren, 1976). The conflict between gown-clad university students and professors, representing the educated elite, versus the blue- and white-collar townsmen was reinvigorated in the 1970s, when new universities mushroomed and admissions exploded, bringing scores of left-wing students to middle-class cities around the world. Some local governments sought to avoid such confrontations by allocating cheap land for the new educational institutions as far away from the town center as possible. Today, the rise of the knowledge economy has changed all that, as knowledge is now perceived as the most valuable resource and learning as the most important process. Town–gown clashes have been replaced by town–gown collaboration (Clavelle, 2001; Cotsones, 2013). Most universities have outreach activities, and most local governments have a collaboration agreement with their local university. Universities increasingly collaborate with government, industry, and civil society (Goddard & Vallance, 2013; McDonald et al., 2023; Schwoerer et al., 2022), and there is a clear trend in the USA and elsewhere towards the formation of university centers and innovation hubs involving external actors in collaborative projects (Kern & Smutko, 2021; Pellegrini & Johnson-Sheehan, 2021; Trencher et al., 2014). The culmination of this development is the formation of city–university partnerships (CUPs) based on the recognition of their mutual resource dependence (Keeler et al., 2016, 2018). This article aims to shed light on the conditions for CUPs to be successful in generating public value outcomes through interorganizational and multi-level collaboration.

CUPs emerge in response to shared ambitions to improve service quality, solve complex societal problems, and create more sustainable and resilient urban environments (Caughman et al., 2020). They aim to provide a platform for collaborative governance and co-creation involving public administrators, researchers, public and private enterprises, and perhaps even lay actors, such as service users, citizens, and civil society organizations (Ansell & Gash, 2018). While CUPs tend to be based on a more or less formal agreement between city officials and the university leadership to collaborate on achieving common goals, the level of ambition and modus operandi tend to differ with the partnerships falling into three different categories (see Caughman et al., 2020; Teitel, 2008): (1) The routine partnership is transactional and consultant-based, characterized by loose exchanges between individuals and limited joint efforts to consolidate and formalize collaboration. The city hosts the university, and the mayor participates in annual university ceremonies, but the ambitions regarding collaboration are limited; (2) Strategic partnerships are based on shared visions and desires to co-create joint solutions, but these partnerships are loose and focused on sector-specific collaboration; for example, in and around the growing number of “university hospitals.” The city is proud to have a university within the city limits and sees it as a valuable resource, although it is far from fully exploited; and (3) The transformative partnership is a formalized partnership based on mission alignment and cross-institutional integration and learning. The city and university see themselves as parts of a “community of fate,” and they collaborate to foster internal and external systemic change. The transformative partnership is more ambitious than the routine and strategic partnerships when it comes to exploiting resource interdependencies to produce innovative public value outcomes, but it presupposes the presence of a strong mutual commitment and a clear sense of common purpose.

A transformative CUP is predicated on horizontal collaboration between two different hierarchical organizations and vertical alignment between the strategic level, where goals and structures are formed; the tactical level, where strategic priorities are turned into actionable agendas; and the operational level, where concrete results are produced and measured. However, since collaboration and alignment is often undermined by institutional inertia, fading attention, cultural differences, and organizational conflicts, recent research has pointed to the need for the exercise of metagovernance, which uses a combination of hands-off framing and institutional design and hands-on leadership, management, and active participation to support collaborative governance in interorganizational networks and partnership (Bailey & Wood, 2017; Jessop, 2002; Peters et al., 2022; Sørensen & Torfing, 2009, 2017). Metagovernance is often exercised by resourceful and legitimate public authorities, but it is important for them to influence the process and outcomes of collaborative governance without reverting too much to hierarchical forms of command and control that will tend to undermine the self-regulatory capacity of interactive arenas (Peters et al., 2022). Of special importance to transformative CUPs, metagovernance may target collaborative governance processes traversing the strategic, tactical, and operational levels of interorganizational partnerships.

Jessop (2004) coined the notion of multilevel metagovernance, referring to the strategic efforts of metagovernors to design, direct, and align collaborative governance systems spanning local, national, and supranational levels. While multilevel metagovernance refers to the metagovernance of the interaction between different jurisdictional levels and CUPs are clearly formed at a single jurisdictional level (the local level of a big city or a metropolitan area). Nevertheless, the concept of multilevel metagovernance provides important insight into the need to create a division of labor between different levels of governance and to let the upper levels of governance strategically frame and direct the actions at the lower levels that play a key role in coordination and problemsolving and may enjoy considerable autonomy.

To further explore the prospect for developing CUPs into local platforms for the co-creation of innovative public value outcomes, such as enhanced social, economic, and environmental sustainability, this article asks: How can metagovernance contribute to aligning different actors and levels in City–University Partnerships aiming to exploit resource interdependencies to enhance public value production? (see Bye et al., 2020). Answering this pertinent question will allow us to better understand how public administrators and researchers can expand their reach, spur knowledge-based governance, and build a joint platform for the co-creation of new solutions to pressing societal problems. The question is answered through an in-depth case study of an ongoing and rather ambitious effort to create a transformative CUP in Trondheim, Norway. In 2017, Trondheim Municipality (TM) and the Norwegian University for Science and Technology (NTNU) launched the University-Municipality Trondheim 3.0 Project (TRD3.0), which aimed to facilitate knowledge sharing, municipal competence development, and the improvement of educational and employment opportunities for students, while simultaneously solving complex societal problems through joint action and co-creating public value together with societal actors and local citizens.¹ The TRD3.0 thus represents a significant systemic innovation aimed at sparking policy innovation as well as service innovation. The project has recently been renewed and expanded, and it will continue its search for the right institutional format in the coming years, eventually developing and disseminating a generic template for the design of transformative CUPs in other university cities.

The article explains, first, the theoretical framework that is based on theories of collaborative governance in networks and partnerships and studies of the role of metagovernance in designing institutional platforms for collaborative governance and in leading and managing collaborative processes through which innovative public value outcomes are created. The theory section is followed by a brief account of the case selection and the mixed methods of data collection. The next section presents the case by explaining the background, drivers, goals, and organizational design of TRD3.0. This serves as a prelude to the presentation of the findings, with a focus on the multilevel metagovernance of the horizontal and vertical relations between actors and levels. The discussion relates the findings to other critical metagovernance experiences and provides some recommendations to practitioners about how to avoid multilevel governance structures, such as the one in the Trondheim CUP, from being stuck in the middle with a tactical level that is disconnected from strategic leadership and has limited operational impact. The conclusion summarizes the argument and sets out a research agenda for future studies of CUPs.

2 THEORIZING METAGOVERNANCE OF COLLABORATION AND CO-CREATION IN PARTNERSHIPS

The study of formal government institutions has given way to the study of formal and informal governance processes through which common objectives are formulated and achieved (Peters et al., 2022; Peters & Pierre, 1998; Torfing et al., 2012). In a world where power is shared and distributed across multiple organizations, sectors, and levels, there is a dire need for cross-boundary collaboration between public and private actors, including for-profit and non-profit organizations and lay actors, such as users, citizens, and local communities (Crosby & Bryson, 2005; Osborne, 2010). While collaborative governance was previously only used when hierarchical and market-based forms of governance had been tried and found wanting, it has gradually become the new normal (Ansell & Gash, 2008; Emerson & Nabatchi, 2015). Hence, it is recognized to produce a “collaborative advantage,” since actors engaged in collaborative interaction can often solve problems and tasks that none of them can solve on their own (Huxham & Vangen, 2013). Collaborative governance may contribute to effective and democratic governance by mobilizing resources and broadening participation (Sørensen & Torfing, 2007). It may also involve a diversity of actors in creative and learning-based problem-solving processes that stimulate innovation, thus creating a sub-species of collaborative governance commonly referred to as co-creation (Ansell & Torfing, 2021a, 2021b; Brandsen et al., 2018; Torfing et al., 2019).

Collaborative governance and co-creation typically take place in interorganizational networks and partnerships. Governance networks can be defined as a relatively sustained interaction between a broad range of interdependent actors from state, market, and civil society who engage in tendentially self-organized collaboration through which complex problems are defined and public value outcomes produced (Klijn & Koppenjan, 2015; Peters et al., 2022; Sørensen & Torfing, 2007). Partnerships also involve public and/or private actors in cooperation with some durability, but they often have a smaller number of actors than networks and are typically based on formal agreements about the purpose of the collaborative interaction, the organizational structure, and procedures for sharing benefits and risks (Koppenjan, 2005). Whereas governance networks are often formed by a broad range of actors aiming to exchange and/or pool their resources to develop and implement joint strategies for problem-solving, partnerships are often formed when two or more actors want to mobilize additional resources and create synergy through collaboration based on transformation and/or the integration of partner organizations. The purpose of partnership formation may include letting the partners act more entrepreneurially while introducing new and more effective or efficient ways of doing things (McQuaid, 2000). Transformative CUPs fit this description very well.

Whether taking place in networks or partnerships, collaborative governance aims to construct a common ground for joint problem-solving based on the constructive management of difference (Gray, 1989). The actors collaborate in defining and transforming a common object (a problem, challenge, ambition, or task) while seeking to overcome emerging conflicts and focusing on jointly formulated objectives. In partnerships aiming for some degree of organizational integration, the collaboration may take place at different organizational levels (Loorbach, 2010). Executive leaders may collaborate at the strategic level to develop joint visions and long-term goals supported by the design and oversight of institutional structures enabling goal attainment. Middle managers may collaborate at the tactical level to transform long-term goals in concrete plans and strategies and facilitate inter-organizational collaboration around specific issues. Finally, relevant and affected staff members may collaborate at the operational level in a range of joint projects in which solutions are developed and implemented, and common tasks are carried out. CUPs tend to combine all three levels of collaborative interaction.

Partnerships are formed between autonomous but mutually dependent actors who choose to achieve goals and solve a particular set of problems and tasks through collaboration supported by purpose-built institutional arrangements. Despite the shared commitment to and interest in collaboration, it is difficult to sustain the horizontal and vertical relations of interaction, as the partners each have different organizational purposes and different rule and resource bases. The chance of successful collaboration and goal attainment increases if the collaborative governance process is properly governed. The governance of collaborative governance is referred to as metagovernance (Bailey & Wood, 2017; Jessop, 2002), and since metagovernance involves the governance of self-governing processes of collaboration in networks and partnerships, it goes without saying that it must be “directive and facilitative” rather than “commanding and controlling.” Heavy-handed metagovernance will tend to scare off the participating actors or generate fierce opposition. The metagovernance tools include the institutional design of platforms and arenas and the exercise of leadership aiming to frame the collaboration, facilitate interaction, and mediate conflicts (Sørensen & Torfing, 2009). The different tools must be used with great care and based on accurate information about the collaborative processes and the emerging problems and challenges. Hence, top-down metagovernance presupposes bottom-up input.

Metagovernance must be exercised by centrally placed actors with sufficient authority to prompt action and mediate conflicts and adequate resources allowing them to lower the transaction costs of collaborating (Peters et al., 2022). The metagovernor must also enjoy organizational backing, which enables the formulation of a precise diagnosis of problems and challenges and monitor progress in joint problemsolving. Public managers typically play the metagovernor role, but they may form a steering group including some key participants (Milward & Provan,c2003). What is less explored, however, is the distributed character of metagovernance in multilayered collaborative governance systems, where leaders, managers, and coordinators at different levels may contribute to the metagovernance of collaborative processes (but see Hubbard et al., 2022). Finally, yet importantly, since some metagovernance decisions are political in nature, elected politicians may team up with the top-level public managers to lend democratic legitimacy to the political metagovernance of networks and partnerships (Sørensen & Torfing, 2016). However, research shows that politicians who are pressed for time and short on specialized knowledge play a marginal role in metagovernance (Koppenjan et al., 2009).

Given the challenge of CUPs such as TRD3.0 to both align well-established organizational actors (and lay actors) in horizontal interaction and align actions at different strategic, tactical, and operational levels along the vertical axis, we would expect lead actors to invest considerable time and energy in metagovernance that draws on a broad array of hands-off metagovernance tools, including the institutional design of a multi-layered yet well-connected collaborative infrastructure, framing of the collaborative process at different levels through storytelling and strategic goal setting, allocation of resources to different action arenas, mechanisms for soliciting feedback from the tactical and operational levels, etc. These hands-off metagovernance tools must be supplemented with hands-on metagovernance tools aiming to lead and manage collaborative interaction through efforts to convene relevant and affected actors, motivate their participation, build trust, construct boundary objects, and mediate conflicts. Direct participation in collaborative processes may also enable metagovernors to influence agendas, decisionmaking premises, and outcomes. Only a skillful combination of hands-off and hands-on metagovernance tools targeting all actors and levels will enable the successful development of a partnership aiming for organizational integration at the strategic, tactical, and operational levels. The challenge for the metagovernors when combining hands-off and hands-on tools will be to avoid excessive metagoverning, which risks pacifying the distributed actors, and insufficient metagoverning, which enhances the risk of mission drift, waning attention, and growing conflicts.

3 METHODS DEPLOYED IN THE EVALUATIVE CASE STUDY

The UniverCity project aims to provide a formative evaluation of the ongoing attempt of TM and NTNU to develop a partnership capable of enhancing coordination, spurring mutual learning, and co-creating innovative public value outcomes. By posing questions and offering critical reflection on the progress, problems, and achievements of TRD3.0, it incorporates elements of the developmental evaluation approach recommended by Patton (2010).

The decision to conduct a single case study is motivated by an interest in providing an in-depth study of a contemporary attempt at using metagovernance to create a successful, transformative CUP. Most case studies of CUPs are from Anglophone countries (Ayoubi & Massoud, 2012; Caughman et al., 2020; Pauer et al., 2020), so to broaden the geographical scope, we decided to study a CUP from Norway that is an affluent western country with a strong collaborative tradition. Here TRD3.0—with its national flagship status—provides a critical case, as it is driven by an ambitious and capable municipality and a highly committed and esteemed university, thus making it most likely that it will succeed in its endeavor. While studying a single most-likely case prevents the generalization of results, it enables a better understanding of the impact on the metagovernance of multi-layered partnerships such as the TRD3.0. It may even facilitate a “reverse Sinatra” inference – “if it doesn't work here, it may not work anywhere” (Levy, 2008) – at least if the project as a whole appears to be well-designed.

The case study is based on mixed methods and draws on multiple data sources, which allows the triangulation of insights and increases the validity of conclusions. The relevant documents, including the initial project plans, a joint memorandum of agreement, relevant steering documents, internal reports, minutes from meetings, etc., were collected and analyzed to better understand the TRD3.0 goals, structure, and activities. About 15 observations of the steering group, committee, and project meetings helped us to get beneath the formal accounts presented in documents and to study the informal interactive processes. Interviews with leaders, middle managers, and staff from TM and NTNU that had been involved at different levels of the TRD3.0 and were selected to enhance diversity were conducted in two stages. In the first stage (2019–2020), interviews with 20 informants from all parts of the TRD3.0 were conducted by the authors to obtain a clear sense of what was going on, the form and content of the interaction, and the drivers and barriers of joint action. These interviews were used to write evaluative interim reports feeding back into the TRD3.0. In the second stage (2021–2022), we conducted focus group interviews with 2–3 participants from each of the joint innovation committees, which we supplemented with interviews of steering group members, coordinators, and project leaders focusing on questions pertaining to the role, tools, and impact of multilevel metagovernance. The interviews were semi-structured, allowing the interviewer to deviate from the interview guide and the informants to tell their story. Some of the interviews were conducted online due to Corona-related restrictions. Finally, evaluative mini-surveys with both closed and open questions were administered twice to municipal and university TRD3.0 participants. The survey questions focused on the motivation to collaborate together with the drivers, barriers, and results. The qualitative answers to the open questions supplemented our interviews, while the quantitative data allowed us to see the prevalence of specific views and opinions.

The material was all coded based on theoretically derived and some emerging codes by different researchers focusing on the purpose, status, and functioning of different collaborative arenas; the participation and interaction in these arenas; the results of the interaction and the emerging problems and obstacles; the role of institutional design and leadership; and the recommendations for future improvement.

One methodological problem is that the data gathered in the early phase had a somewhat broader focus than this article and mainly contributed to understanding the nature and ambitions of the CUP. However, the data collected in the final phase was guided by the research question that we aim to answer here and which provided the rich and multifaceted material allowing us to draw solid conclusions. Another problem is that some of the researchers evaluating the TRD3.0 (and co-authoring this article) were also involved in spearheading it. To solve the problem of potentially conflicting interests and lack of critical distance, an external expert evaluator from a different country was included as co-evaluator and co-author.

4 TRD3.0—BACKGROUND, DRIVERS, GOALS, AND ORGANIZATION

TM and NTNU jointly initiated the TRD3.0. The two founding organizations are large, capable, and strongly committed to creating a transformative CUP. TM has 15,000 employees and NTNU 7400 employers and their leaders expressed their strong dedication to TRD3.0 during interviews. Moreover, the antecedent conditions for constructing an institutional platform for the coordination, integration, and co-creation of public value are favorable. TM and NTNU have a lengthy record of working together, even if the collaboration has been somewhat uncoordinated, relatively informal, borne by individuals, and often with NTNU as the proactive part. Moreover, the mission statements of both organizations tend to link sustainable societal transformation with knowledge generation and collaboration, thereby producing a notable ideational congruence, although TM is more oriented towards meeting local needs whereas NTNU is more oriented towards international quality and excellence. Finally, both TM and NTNU are credible conveners capable of attracting other relevant and affected actors and facilitating co-created problem-solving. They have demonstrated their convener power many times in the past when building governance and research networks.

Aside from TM and NTNU sharing the ambition to achieve the UN Sustainable Development Goals based on collaborative knowledge generation, both steering committee and project group meetings revealed that the two organizations have rather different motives to deepen and institutionalize their collaboration: TM wants to play a larger role in joint R&D projects and develop the professional competences of its employees through research-based training and education, whereas NTNU wants to recruit more students for mid-career education and improve and enable the practical training of students based on explorative projects and internships. The latter is also aware of how future access to research funding relies on the ability to involve public and private stakeholders. Both national and international (EU horizon programmes) research funding depends on partnering with some practical cases, and the municipality provides potential access to such cases. While both organizations are doing relatively well in terms of research partnerships, there is great potential for improvement.

TRD3.0 aims to promote deeper inter-organizational collaboration and integration between TM and NTNU while also spurring the interaction with citizens and local stakeholders to produce public value for the involved organizations and society at large. At a concrete level, the institutional integration involves harmonizing the annual planning cycles of the two main partner organizations, the development of models for sharing or exchanging personnel, and use of the municipality as a laboratory for research and teaching. It also involves the formation of a multi-layered organizational design consisting of a joint steering group and five joint innovation committees, each with a portfolio of project groups that may or may not include the participation of external societal actors. The steering group, the innovation committees, and project groups operate at different levels undertaking strategic, tactical, and operational functions and tasks, respectively. The three organizational levels are supposedly joined through a process of top-down strategic planning whereby the steering group sets the direction, the innovation committees translate overall goals into a set of priorities and proposed activities, and the project groups develop and implement solutions. TRD3.0 has also launched four cross-cutting initiatives relating to different organizational missions, and the goal is to develop a generic model for formative CUPs that can be translated into a national standard that will enable universities and municipalities to strengthen their collaboration, obtain state recognition and receive a state grant to help to finance the exploitation of local synergies.

The inter-organizational collaboration in TRD3.0 is supported by digital platforms and tools based on open standards. The digital tools facilitate information gathering, data sharing, online meetings, etc. The financial support for TRD3.0 comes from the two main partner organizations, which split the costs evenly while working proactively to secure external funding via regional, national, and European research grant applications. The initial budget for TRD3.0 for the period 2017–2021 was NOK 30 m (approx. €2.5 m). The project started in 2018 with a formal agreement endorsed by the City Council and signed by both parties.

5 EMPIRICAL FINDINGS REGARDING THE METAGOVERNANCE OF THE TRD3.0

Fostering inter-organizational collaboration at multiple levels is a difficult task, even when the conditions are ripe, and the partners are capable and committed. Hence, earlier research has identified several barriers for the TRD3.0 (Torfing et al., 2021). First, in big, complex bureaucratic organizations such as TM and NTNU,² new strategic priorities typically take a long time to trickle down to the tactical and operation levels, especially if there is no deliberate attempt to secure ex ante support for the middle managers and frontline staff. Second, as public organizations with a broad range of tasks, both organizations are pressed for time, which makes it difficult to prioritize something that is not a primary task, albeit the existence of a formal agreement legitimizes spending time on collaborative efforts. Third, while TM is eager to enhance collaboration and integration to facilitate research-based training and service improvement and to create innovative responses to complex societal problems, NTNU does not face the same pressure to deliver better service and public value outcomes, and the researchers cherish their autonomy and research freedom, which means that they appear slow and foot-dragging with respect to investing in new forms of collaboration. Discrepancies in perceived commitment tend to undermine trust, which in turn renders inter-organizational communication and collaboration difficult (Covey & Merrill, 2006).

These and other barriers to the formation of a transformative CUP such as the TRD3.0 are exactly what call for the exercise of metagovernance. This implies creating, evaluating, and revising institutional designs that support horizontal collaboration between actors and help to align the different levels along the vertical axis. The key word in such hands-off metagovernance is connectivity: creating arenas that connect relevant and affected actors and connecting goals, initiatives, and activities at different levels. Metagovernance also involves managerial efforts to frame the collaborative endeavor, facilitate interaction, mediate emerging conflicts, and participate in the collaborative process to influence decision-making. These forms of hands-on metagovernance must be undertaken by experienced, competent leaders who have a clear assignment and are well connected across the different levels of the interorganizational collaboration.

Let us examine the exercise of hands-off and hands-on metagovernance more closely at the strategic, tactical, and operational levels while looking for the interconnections. Deploying this analytical framework in empirical analysis will shed light on the problems of forming transformative CUPs under favorable conditions.

6 DISCUSSION: IMPROVING MULTILEVEL METAGOVERNANCE TO CREATE FLOURISHING TRANSFORMATIVE CUPs?

As the interest in strengthening town–gown collaboration increases and matures, we must evaluate empirical models and explore and assess the role of metagovernance in connecting actors and levels. Our study of the transformative CUP in Trondheim, Norway, provides new knowledge about how cities and universities may organize their endeavors to strengthen and deepen collaboration and perhaps even co-create sustainable public value solutions. While the TRD3.0 is an ambitious project that has taken bold steps towards strategic and institutional integration, enhanced the commitment to further collaboration, and generated a few tangible results, we find that the metagovernance of the collaborative processes at the strategic, tactical, and operational levels is deficient, as it results in the TRD3.0 being stuck in the middle, with the tactical level largely decoupled from the strategic and operational levels.

This finding is important because it sheds new light on the role of metagovernance, which is rarely studied at the local level and in the field of town–gown collaboration and seldom pays attention to the challenge of metagoverning multi-layered systems of interorganizational collaboration aiming to connect strategic, tactical, and operational levels of governance. Extant studies aiming to assess CUPs mainly focus on the integration of actors and pay scant attention to the integration of the different levels of interorganizational collaboration (see Caughman et al., 2020). The present study compensates for this benign neglect by showing how difficult it is to metagovern multi-layered collaborative governance processes involving executive leaders, middle managers, and staff from different organizations. Public managers are well-trained to lead and manage their own employees using a combination of sticks, carrots, and sermons, whereas they have less experience with leading collaborative governance processes involving actors from different organizations and operating on different levels.

Our main finding also helps us to realize how, although the creation of transformative CUPs sounds like a good idea, making them work in practice is difficult, even when the partners are committed and there is positive past experience with collaboration. While we are close to being able to make the “reverse Sinatra” inference claiming that “if it does not work here, it may not work anywhere,” we want to be cautious, in part because the TRD3.0 remains in its infancy, has been negatively affected by the COVID-19 lockdowns, and is strongly committed to learning from the initiation and consolidation phases to ensure better horizontal and vertical alignment. Another reason for caution is that TRD3.0, despite its lofty ambitions, is based on joint objectives that are vague and unable to compete with the daily chores.

This latter observation brings us to the key question of what the TRD3.0 (and perhaps other transformative CUPs) can learn from our study. Five recommendations can be made based on the above analysis. First, when forming joint organizational units such as the steering group, the innovation committees, and the project groups, it is important to select members from the respective partner organizations who are committed to the collaborative endeavor and have the power and authority to “make things happen” (see Bryson et al., 2002). At the operational level, the stakeholder analysis must be expanded to include relevant and affected societal actors who can contribute to the co-creation of the solutions pursued by the CUP. More importantly, the members of the collaborative arenas must engage in relational coordination based on high-intensity communication to build trust and define and pursue common agendas (Gittell, 2000). The latter requires frequent meetings with clear agendas and a structured process that ensures progression from explorative deliberation based on external input and feedback to strategic decision-making and a focus on execution and evaluation.

Second, while it is important to have a central steering group that can critically evaluate and constructively adapt the institutional design of the CUP, the steering group must have direct, tactical-level interaction with the middle managers to avoid decoupling. Centrally placed coordinators may provide administrative support for the steering group, but their effort to act as intermediaries linking the strategic and tactical levels is bound to fail because they have no leverage, which makes conflicts with the leaders at the tactical level a story foretold. Instead, regular interaction between the steering group and tactical-level leaders is necessary to ensure the alignment of goals and visions and to empower the tactical leaders to make clear priorities.

Third, while there is a sound metagovernance focus on institutional design and the need for continuous adjustments, much more focus is necessary on framing the collaboration at the different levels of the CUP by means of repeatedly discussing, gradually clarifying, and constantly translating the vision, the goals, the values, and the joint priorities of the CUP. The dissemination of stories about small wins and major achievements would also be helpful. Framing is important to provide the leaders on all levels with a sense of meaning and direction, but also to mold the views, perceptions, and objectives of the participants that are not set in stone but likely to be transformed through interaction and evolve in conjunction with organizational and environmental changes (Siegel, 2008; Wyborn & Bixler, 2013). In the future, new joint storytelling about the CUP organizational mandate may reduce the discrepancy between the local focus on the municipality and the international focus of most researchers by spreading the message that scientific studies of local processes and results through collaboration with the municipality can feed into large EU projects and international publications.

Fourth, to further integrate the partner organizations and align activities at the strategic, tactical, and operational levels of the CUP, it might be constructive to create an integrated annual planning cycle. The “annual CUP governance wheel” could begin with a short phase of joint agenda setting at the strategic level based on an internal and external analysis of governance demands and input from the tactical and operational levels (1½ months); based on goals and priorities resulting from the agenda-setting phase, the tactical level could then explore the need for adjustments to the project portfolio, with a special emphasis on the need to form new projects and convene the “right” actors (1½ months); after that, the collaborative R&D projects should have good time to co-create solutions focusing on prototype development and testing, which spurs dialogue and rapid learning with frequent reporting back to the tactical and strategic levels (8 months); this procedure would eventually leave time for evaluation and the preparation of the next roll.

Finally, while there is initially good reason to see the creation of transformative CUP as a joint project connecting the partner organizations in an integrative manner, it is important to consider when to end the project status and to make the TRD3.0 a permanent arrangement. The transition from being an external collaboration project to an imbedded mechanism for spurring collaborative R&D may grant the TRD3.0 personnel more authority and may involve more people from the operational level in the innovation committees.

Our analysis points to the need for more rather than less metagovernance, but we must caution ourselves since there is always a dual risk of metagoverning too little (understeering) and too much (oversteering). In the case of the TRD3.0, however, the leadership aspect of metagovernance has clearly been under-prioritized, leaving considerable room for improvement. Most importantly, however, our analysis points to the need for metagovernors to pay more attention to connecting the strategic, tactical, and operational levels of the collaborative endeavor.

7 CONCLUSION AND FUTURE AGENDA

In turbulent times, when new complex problems and challenges prompt the public sector to innovate its system of governance, experimentation provides an important tool for bold and ambitious local governments. Generative experiments, defined as the process of generating and iteratively refining a solution concept (Ansell & Bartenberger, 2016), are helpful for developing new forms of governance that work in practice. The TRD3.0 provides an ambitious and promising generative experiment aiming to create a transformative CUP, and as a part of the ongoing effort to evaluate and even redesign the local efforts, we have asked how multilevel metagovernance can contribute to connecting actors and actions at multiple levels. The theoretical answer to this question is that metagovernors may use a combination of hands-off and hands-on tools to foster horizontal collaboration at the strategic, tactical, and operational levels while linking the levels as a part of a comprehensive governance system. The empirical answer is that the multilevel metagovernance must combine hands-off and hands-on metagovernance tools while paying special attention to connecting activities across different levels. Regarding hands-off metagovernance, there has been a clear effort to design and re-design the TRD3.0 but a failure to properly frame the collaboration between the actors, connect the three levels of the CUP, and allocate resources for collaborative activities at the tactical and operational levels. Regarding hands-on metagovernance, we find that despite laudable efforts to improvise, there has been relatively weak leadership at the strategic level (by the coordinators), at the tactical level (the chairs of the innovation committees), and at the operational level (the project leaders), which has meant failure to connect the activities at the different levels.

The result of these deficiencies is that the tactical level is stuck in the middle and fails to translate strategic priorities into operational activities; and that collaboration is strongest at the operational level, where it continues to be nurtured by personal contacts and pre-existing connections rather than the efforts of the innovation committees and the TRD3.0 project leaders. This negative assessment neither takes away from the manifold achievements of the TRD3.0 nor prevents the involved actors from reaping the promised fruits of a transformative CUP in the future. However, the critical diagnosis must be followed by learning-based attempts to rethink and boost the exercise of multilevel metagovernance. Above, we have provided some ideas for how to improve the metagovernance efforts.

The limitation of this study is obvious, as it merely explores the endeavor to create a transformative CUP based on a single case study. The formation of transformative CUPs remains a ground-breaking and understudied endeavor, however, and we have only begun to conceptualize their form and functioning and the efforts to sustain the horizontal and vertical connections through metagovernance. Further studies should engage in comparative analysis based on a refined understanding of the use of different hands-off and hands-on metagovernance tools, their practical application, and their impact on goal achievement.

Local governments and universities stand to gain much from following in the footsteps of the TRD3.0, and comparative case studies may help to develop nationally transferable templates for transformative CUPs and instructions for metagovernance that ensure success. Going down this road may support the realization of the “civic university” vision, which takes a holistic approach to engagement with wider society to create an impact that transcends academia (Goddard et al., 2016).

FUNDING INFORMATION

The UniverCity project has received funding from the Norwegian Research Council (Grant number: 295802).

CONFLICT OF INTEREST STATEMENT

There is no conflict of interest in relation to the publication of this paper.