Collaborative design of governance instruments in higher education

ABSTRACT

There is considerable focus on finding effective governance approaches. This article examines experiences with using more collaborative approaches to developing new governance instruments in the context of higher education. The specific empirical case focuses on the introduction of multi-annual performance agreements between the ministry and the higher education institutions in Norway. Specific focus is on the pilot process, during which the instrument was developed in collaboration between the ministry and representatives from the higher education institutions. The article takes a starting point in a contingency model for collaborative governance, and examines starting conditions, institutional design and leadership models, and the characteristics of the collaborative process itself, including the role of face-to-face dialogue with good faith negotiation, a trust-building phase, a commitment process, followed by a shared understanding of the problems, values and goals, and intermediate outcomes. The data consist of approximately 50 qualitative interviews with representatives from the ministry and the higher education institutions and a range of document data. The paper analyses how the collaborative design process in Norway shaped policy outputs, and highlights critical factors of the process, including the consequences of the temporal sequencing of the process, the interaction of the new governance instrument with the existing instrument mix, and the role of various other concurrent policy developments. The analysis also points out that while collaborative approaches can provide ways to strengthen legitimacy of governing, they also involve a number of dilemmas and challenges, and therefore require careful design.

KEYWORDS:

• Higher education
• governance
• collaborative governance
• performance contracts
• Norway
• performance agreements

Introduction

Governments are continuously searching for ‘what works’ in higher education governance. One suggested means is to a larger extent involve the stakeholders, expecting that this would strengthen legitimacy and reduce implementation costs, an approach that has been labelled ‘collaborative governance’. Collaborative governance includes a heterogenous set of practices, but principally implies involvement of various kinds of actors in formal, consensus-oriented decision-making processes concerning policy development or implementation (Ansell and Gash 2008, 544, Emerson et al. 2012). In this article, the focus is on examining collaborative approaches in the context of governing higher education, the dilemmas that emerge in the design process, and the potential consequences for policy outcomes.

The empirical setting in this article is the introduction of multi-annual performance contracts in higher education in Norway (labelled as ‘development agreements’). Performance contracts represent a policy solution that has in the last 20 years been employed in a number of countries in Europe and North America, associated with an aim to strengthen differentiation and enhance quality (Benneworth et al. 2011; De Boer et al. 2015; Gornitzka et al. 2004; Jongbloed and de Boer 2020). The stated assumptions that underpin the introduction of contracts are that system diversity in higher education is desirable, and that a one-size-fits-all approach for performance assessment creates perverse incentives for institutions to become more similar.

In the Norwegian context, the contracts were introduced through a three-stage pilot from 2016. Dialogue and consensus-orientation between stakeholders has been an important characteristic of higher education policymaking in Norway (Bleiklie and Michelsen 2012), yet the pilot concerning the development of the contracts introduced a number of new ideas, as the sector was involved as co-constructors of the agreements as a governance instrument. Key questions for the article are: How were the contracts developed? What kind of role did the collaborative elements have for policy outputs? A key contribution of the article is to examine how involvement of the sector in policy design process can create new challenges. The analysis focuses on the policy development process, and highlights the consequences of temporal sequencing of the process, the varying and rather consequential effects of the very open-ended approach from the ministry, the interaction of the new governance instrument with existing instrument ‘mix’, and the role of various other concurrent policy developments. As such, the analysis points out a number of dilemmas and challenges, both around the introduction of performance contracts, as well as those potentially inherent in using collaborative approaches in higher education.

Collaborative governance

Collaborative governance has developed alongside the often-quoted shift from government to governance. The shift implied focus on steering rather than top-down hierarchy, highlighting the need for coordination and cooperation in policy development and implementation. This shift was associated with governance failures from traditional implementation processes and challenges associated with finding solutions to wicked problems (Torfing and Ansell 2017). The growth in both institutional capacity and development of more complex knowledge requires new modes of governance. In other words, as ‘knowledge becomes increasingly specialized and distributed and as institutional infrastructures become more complex and interdependent, the demand for collaboration increases’ (Ansell and Gash 2008, 544). These new modes of governance have acquired a number of different labels, with partially overlapping definitions. Concepts such as participatory governance (Fischer 2012), network governance (Provan and Kenis 2007), interactive governance (Torfing and Triantafillou 2011) all point towards modes of governing where traditional top-down modes of a sovereign state would be to a larger extent be replaced with more interactive, networked, and collaborative modes.

This heterogeneity of concepts also highlights a degree of ambiguity concerning what constitutes collaborative governance. A key element in various definitions is emphasised on the role of interactions with different kinds of partners and stakeholders during the whole policy process, by using a diversity of methods, such as dialogue, deliberations, consultations, negotiations, and more (Bingham 2011). Different definitions for collaborative governance have been presented. For example, Ansell and Gash (2008, 544) present a more restrictive definition where focus is on involvement of non-state actors, and focus is on formal, purposeful processes of policy development and implementation. Others, e.g. Emerson et al (2012, 2) include definitions where the role of actors is broader, including ‘[engaging] people constructively across the boundaries of public agencies, levels of government, and/or the public, private and civic spheres’. These variations of definitions suggest that collaborative governance takes many forms, and definitions vary concerning relevant actors and arenas. The central aspect of those is the focus on meaningful collaboration which extends regular consultations. It is this collaboration that is in focus here, as the ministry engaged in an in-depth collaboration process together with representatives for the institutions.

Emphasising ‘collaboration’ in governance arrangements implies a distinct mode of interaction, distinct from other frequently used modes such as cooperation and coordination. These represent a continuum of increased depth and scope of interaction. Cooperation implies a choice to work together within existing frameworks yet maintaining individual goals and interests; coordination implies that formalised links are established between participants in this process. Collaboration represents the most integrative mode of joint activities, characterised by ‘participants who work together to pursue complex goals based on shared interests and a collective responsibility for interconnected tasks which cannot be accomplished individually’ (McNamara 2012, 391). An emphasis on ‘collaboration’ in collaborative governance thus means that not only is there interaction between various actors involved, but interaction also has a shared purpose in its consensus-orientation.

While consensus-orientation and stakeholder involvement may sound like an appealing approach, does it deliver? Ansell and Gash (2008) reviewed existing literature to develop a contingency model and suggest that this is dependent on starting conditions, institutional design and leadership models, and the characteristics of the collaborative process itself. While Ansell and Gash originally developed this for a narrower definition of collaborative governance, the specific elements are also suitable for this analysis, as they represent a structured means to deconstruct a policy development process. Among the starting conditions, Ansell and Gash suggest that power-resource-knowledge asymmetries, incentives for cooperation, and prehistory of cooperation or conflict as the three main conditions influencing collaborative processes. The imbalances mean that the capacity of various actors to participate or even enter the arena may vary substantially. The latter may be due to organisational infrastructure as well as knowledge and expertise. Having this imbalance in mind, as participation in collaborative governance arrangements would largely be voluntary, there is a question of why stakeholders would engage in this process. A key rationale is that actors expect clear outcomes from their participation, and that this collaboration can provide value-added as compared to other means. These expectations are also related to the third element, which is the prehistory and preceding levels of trust between the various actors. The more trust, the higher the likelihood for a virtuous cycle of collaboration. Ansell and Gash also find that leadership appears to be a critical factor in bringing participants to the collaboration and assist with reaching consensus among participants. It is particularly the leaders’ ability to safeguard the process that is being emphasised. This is particularly important with unfavourable starting conditions (e.g. power imbalances, distrust, etc.). Similarly, institutional design of the collaborative process matters, that is, the basic rules for the design of the collaborative process. This includes decisions about access (inclusive vs exclusive processes), process transparency, as well as rules about expected degree of consensus and deadlines. The characteristics of the collaborative process matter as well. While a number of different models exist, Ansell and Gash emphasise the cyclical and iterative nature of the process. The process is dependent on communication and face-to-face dialogue between participants. Collaborative processes require trust-building and commitment to the process. During the process, a shared understanding should be developed and preferably, some intermediate outcomes could be observed that could re-emphasise a virtuous circle of trust and communication. Overall, Ansell and Gash suggest that time, trust, and interdependence are critical factors in collaborative governance. This contingency model is used as a lens to analyse empirical data in this article.

Governance of higher education and the introduction of performance contracts

Higher education governance in Europe has since the last decades of the twentieth century been undergoing a substantial transformation. Being inspired by the US higher education, stronger market elements have been introduced (Gornitzka and Maassen 2000). The ‘rise of the evaluative state’ introduced indicators, criteria, and targets as essential means of governance arrangements (Neave 1988). Application of ‘new’ governance models profoundly changed governing of all aspects of higher education, introducing emphasis on performance funding, formal external quality assurance of education, new forms of leadership and management, to name a few. Many of these changes also strengthened the element of quantification in governance. Nevertheless, as the use of indicators and targets became proliferated, this did not imply any less control, often it merely suggested control by other means (Gornitzka and Olsen 2006). By building a tight web of various measures to assess performance and hold institutions accountable, the ‘new’ governance modes allowed for sustained control, while simultaneously also increasing the formal autonomy of the higher education institutions. The rise of formal institutional autonomy in higher education can thus be found in the ‘shadow of hierarchy’, where control and steering is merely enacted by alternative means (de Boer and Enders 2017). This means that while higher education governance in Europe includes a larger share of market elements, the overall ‘steering load’ in many countries has not necessarily been reduced, not least as overall governance trends in recent years also point towards more pronounced emphasis on re-regulation (Capano and Pritoni 2020).

System differentiation frequently seems to be a desired goal in many countries, challenged by uniformity of environmental demands (Van Vught 2008). When funding is based on specific kinds of performance-based formula, this promotes a specific model or template of a successful institution (Jongbloed 2011, 188). As a means to address these challenges, a number of European countries have explored various forms of performance contracts (De Boer et al. 2015; Jongbloed 2011; Jongbloed et al. 2018) that reward higher education institutions ‘on the basis of the performance that an institution expects to deliver in the (near) future’ (De Boer et al. 2015, 8). Focus is shifted towards future goal attainment as a core element of steering (and funding). Goals and expectations between parties become formalised (Gornitzka et al. 2004), and enhanced focus is put on accountability (Jongbloed et al. 2018). Yet, such agreements also represent a highly heterogeneous approach to governing. De Boer et al. (2015) found a large variety of labels being used: compacts, target agreements, outcome agreements, and development plans. Moreover, they differentiate between ‘hard’ and ‘soft’ contracts – depending on whether performance assessment would be based on ‘prescription of a certain outcome’ or ‘the efforts an actor reasonably has to make’; and whether the contracts have funding associated with them (performance contracts) or not (a form of letter of intent), among else (12–13). The contracts vary substantially in the degree to which the institutions are invited to participate in determining the goals that would apply for them. While these studies have examined how contracts have been implemented in various systems, fewer studies have examined how contracts as instruments are being developed, which is the aim of this article.

Data and methods

The analysis presents a case study of introducing multi-annual ‘development agreements’ in Norway from 2016 and onwards, in three consecutive rounds; first as a pilot to five institutions, then another five, followed by the remaining 11 institutions in the sector. The data were collected in 2020, when the first pilot round was about to be complete and decisions about the future of the agreements were about to be made.

Document material includes the agreements themselves, as well as various meeting notes and copies of presentations held during the process. Interview data consists of 56 interviews in total, each about 45–60 min. Data collection procedures (incl. information, consent, data storage) were approved by SIKT.¹ Eight of these were conducted at the Ministry of Education and Research, representing informants with leadership functions and those in regular managerial positions. These interviews explored the process of introducing the agreements, the roles and views of various actors in the process, how the agreements fit with existing governance instruments and their insights about the outcomes at the point. In addition, interviews at ten higher education institutions were conducted. These ten represented the ten institutions involved in the first two rounds of the pilot. Informants at the institutions included academic and administrative top leadership (e.g. rectors, vice/pro rectors, directors, and equivalent), board members/leaders, as well as some informants on the level of middle management (e.g. faculty/unit leadership). Of the 48 interviews, there were about 4–6 interviews per institution. The interviews had its main focus on the involvement of the institutions in the process of developing the contracts, as well as various aspects about their local implementation processes. All of the 56 interviews were recorded and transcribed fully, and then analysed based on pre-existing codes that systematised the starting conditions, institutional design and leadership, as well as characteristics of the collaborative process. From the 48 interviews at the institutional level, only data concerning the development of the agreements as a steering instrument have been used in this article.

Background and context for the study

Governance arrangements in Norway can be characterised as steering based on goals and results. Since 2015, the overall governing of the sector was based on four overarching sectoral goals (sektormål), each with quantitative parameters. There was an expectation that institutions would define their own institutional goals within this framework (virksomhetsmål), taking into account their specific profile. Institutions report annually both on the sectoral goals and their own goals, in addition to other reporting. Key critiques have been that there has been an abundance of goals and targets, and that reporting to different and often not sufficiently well coordinated public bodies leads to inconsistency (Lyby et al. 2020). A performance-based component in the funding system was introduced in 2003: about 30% of the overall budget was allocated based on quantitative indicators for education and research (e.g. study point production, publication points). Some represented a zero-sum game (e.g. publication points), others are open in that higher results would lead to more for everyone (e.g. study point production). The funding system was evaluated several times, and was until recently mostly adjusted over time rather than fundamentally changed.

An important concern in Norwegian higher education policy, given its reasonably small size and system structure, is the notion of division of labour and differentiation. In 2015, the funding system was evaluated by an external expert committee, suggesting supplementing the funding system with an instrument that would also reward differentiation and strengthen institutional profiling – performance contracts (Hægeland et al. 2015). During the public hearing, the idea of contracts was received with curiosity, but the specific model that the expert group had suggested received scepticism. The ministry didn’t push through an unpopular version of contracts and instead proposed a pilot initiative, with no funding attached to the contracts at the time. It was announced that funding may/shall be attached at a later stage.

Collaborative elements of the policy development process

In this section, some of the collaborative elements of the process will be explored in more detail, taking a starting point in Ansell and Gash’s (2008) set of contingencies.

Starting conditions: trust in the shadow of hierarchy

Ansell and Gash (2008) suggest three key elements as starting conditions: prehistory of cooperation/conflict, the various power-resource-knowledge asymmetries, and incentives for cooperation. At the time of introducing the pilot, Norwegian higher education governance was characterised by its high degree of trust and involvement, and consensus-oriented approach to policy development and governance (Maassen and Stensaker 2021; Stensaker and Michelsen 2012). Reform processes have been characterised by extensive consultations and dialogue during agenda setting and policy formulation, and generally low conflict. The ministry had a tendency to back off controversial and unwanted proposals:

There really is a focus on being collaborators and respect institutional autonomy. This is of course related to how the ministry is developing policies and obtains feedback. Every time we start with a new White Paper we have extensive meetings and hearing rounds and a lot of dialogue. So, when we finally publish the White Paper there are no surprises, and this becomes part of the implementation as well. So, much of our policymaking is dialogue based. (ministry)

Dialogue incentivised engagement – the ministry was looking for a way to adapt the instrument to solve the issues of uniform governance pressures and assure implementation, the institutions had a self-interest to avoid or shape a potentially non-desirable governance approach. Despite dialogue-orientation, asymmetries nevertheless exist. There are power differences between the various higher education institutions as the institutions vary substantially in size and output. Moreover, these are publicly owned institutions – even though the ministry engages in dialogue-based governing and policymaking, ‘eventually someone makes a decision’, which means that there is also an awareness of an imbalance. The dialogue is thus granted by the ministry and is conducted at their privilege. One of the informants explicitly referred to the notion of ‘shadow of hierarchy’ in the interviews, to describe the dynamics of dialogue between the ministry and the sector. Nevertheless, an overall core theme is emphasis on trust and a tradition for collaboration. As such, the process had reasonably productive starting conditions.

Design and leadership: openness with constraints

The second set of contingencies refers to the specific framing conditions of the collaborative process. These concern the design of the process: inclusiveness, exclusiveness, clarity of ground rules and process transparency (Andranovich 1995; Ansell and Gash 2008); and leadership as a means to engage actors and empower them (Ansell and Gash 2008).

The pilot was designed in three separate rounds. The participation by the institutions was voluntary, but the temporal sequencing created issues with participatory inclusiveness as options to participate were stepwise and thus barred access to some institutions who would have wanted to join the first round. Rather than including all interested parties, the ministry prioritised including institutions with different profiles, in this manner representing a rather curated and diverse group. A diversity of participants has elsewhere been found to be a benefit for collaborative processes (Andranovich 1995). This was deliberate design: ‘We considered this and then we contacted with a few institutions with different characteristics to start with, and asked whether they wanted to join this process. And then all five of them said yes’ (ministry). This was explained with uncertainty about the contracts and how they would function, so starting with all 21 institutions was seen as too complicated. While small group facilitated discussions, there were trade-offs as division of labour in the whole sector was difficult to discuss. Importantly, while the ministry established broad ground rules for the process, the shape and form of the instrument itself was left open – explicitly inviting the sector into the collaborative process. There was also a high degree of transparency over the process in general. As such, it would also seem a number of these contingencies were also in place.

Leadership has an integral function in assuring success in complex collaborative processes (Crosby and Bryson 2010). In the case of the contracts, the process was initially embedded in the top political leadership of the ministry. A minister or state secretary participated in the plenary meetings, but the main responsibility for negotiations and management was delegated to the ministry bureaucracy. According to some informants, there was a somewhat diluted interest after the political leadership in the ministry changed during the process, which may suggest weakened leadership.

Collaborative process: open but ambiguous dialogue

In the model proposed by Ansell and Gash (2008), the collaborative process is described as a circular and iterative activity with specific elements: face-to-face dialogue with good faith negotiation, a trust-building phase, a commitment process where mutual interests are being agreed upon, followed by a shared understanding of the problems, values and goals, and some indications of immediate outcomes.

Face-to-face dialogue. The three rounds of the pilot process each consisted of several plenary meetings with leadership from the involved institutions. The first was held in February 2016, where the institutional framework for the process was set. Ministry representatives presented the ambitions and aims for the contracts and an outline of the pilot process, but essentially only outlining some basic ground rules. After this, the institutions were asked to identify three thematic areas for goals. In April 2016, another plenary meeting was held to discuss the drafts. The contracts were also discussed during the regular annual steering dialogue meeting (etatstyringsmøte) later in the spring. After this, the contracts were followed up locally at the institutions. In the autumn, a more detailed bilateral discussion followed between the institutions and the ministry. By late autumn, the contracts were finished and included in the state budget – the contracts were added to the annual state budget allocation documents (tildelingsbrev) of the institutions. All three stages followed a similar approach, with the exception that successful examples from previous years were used in 2017 and 2018. Both informants from the ministry and the sector highlight the role of the shared and sufficiently small face-to-face meetings as an important element:

This has had positive effects in that we can use the experiences, and we have had time to discuss with the institutions. It is easier to talk in a room where there are not too many people present, it does something to the discussion. (ministry)

The ministry came to the process only with some open and rather overarching ideas of how contracts could look like (not too many goals, that they should be measurable, etc.) and certain no-go areas (commitment of new funding from the ministry). This strategy was intentional: ‘we wanted the institutions to be part of this discussion, so we did not set any specific template’ (ministry). Most informants described a sense of good faith in the discussions:

We felt that the contracts came to be as we went along, and felt the ministry was very attentive to how we would imagine that they should look like. We really experienced that this was a contract that two parties were developing in collaboration. (HE sector)

While many would also refer to the fact that there is a power imbalance, some also suggest the discussions this time it didn’t always feel like this: ‘the dialogue here really was not about instructing the institutions, we wanted to have an open dialogue. The institutions also said this was important for them, so we respected this’ (ministry). As such, the ministry did not enforce its authority when debating the overall shape of the contracts. The only instance where the ministry did to some extent retort to hierarchy was when they were suggesting adding some goals, e.g. teacher education as this was a high political priority at the time. In general, hierarchy seemed a more internalised constraint:

I do feel the [institutions] are quite attentive to what we say, but I think we were also quite strong in saying that this is something you have to own, and I do feel the discussions were real. But still, however you want to spin it, we are the owners (…) and of course I understand they follow what we say. I did not feel this was an issue. But of course, looking externally, the dynamics is there. (ministry)

This generally open approach along with a relative lack of enforcing authority seemed to convince the participants of good faith in the process. Moreover, it also invited the institutions to a genuine discussion, thus decreasing potential attempts of gaming among participants.

Trust building. As a reasonably high degree of trust was one of the starting conditions, this dimensions primarily concerned with trust concerning this particular process. Here, the open approach was an important mechanism. One side of this was institutions being open about their own processes and mutual learning: ‘it was four-five institutions present, the discussions were really good. They were really open with each other, about how they had developed their contracts’ (ministry). Moreover, the importance of cumulative learning was emphasised:

We got them to challenge each other, and we also included someone who had gone through the process the year before and could talk about it. So, this was an attempt to engage in learning, and especially in the next round where they challenged each other. (ministry)

Overall, the general good faith obtained in the face-to-face dialogue reinforced existing trust in the dialogue. Nevertheless, the worry about funding potentially being attached to the contracts was lurking in the background: ‘of course one then starts to think, this will be a zero-sum game, and someone, probably, will lose and some will benefit’ (HE sector). During the initial phases of the pilot, the question of funding remained undecided.

Commitment to the process. While openness was on the one hand seen as an invitation for the sector to also be involved, paradoxically this also led to some institutions become uncertain about the purpose and how committed the ministry really was. According to some, this was further reinforced after a shift in political leadership at the ministry. While most institutional representatives were rather enthusiastic about the opportunity of more differentiation and to be seen for ‘who they were’, some did not see the added value, and consequently had low commitment:

It is not clear how the ministry would support us in achieving those goals. Therefore, it really is yet another document that the ministry can use to say that look, here you have not succeeded, here you managed ok, and here you maybe succeeded. But we already have the steering dialogue for that, along with the existing indicator structure. (HE sector)

Some of the institutions had been undergoing merger processes, which meant that the introduction of the contract did not always sit well with the rest of their internal merger timeline – it was difficult to establish a contract when there is not yet an institutional strategy: ‘We were really overwhelmed, (…) very many things had to get into place at the same time. But of course, this was also a time where everything was up for debate’ (HE sector). This suggests that the radical changes also presented a window of opportunity, while it may also have weakened their overall commitment.

Overall, there are rather varied preferences about the future of the contracts, and the kind of instrument they should become, whether there should be money involved, and if so where that money should come from.

Shared understanding. When members of the process are committed, they are more likely to develop a shared understanding (Ansell and Gash 2008). This implies both an epistemic and normative dimension, including a shared definition of what the policy problem is, a clear mission, and common values that underpin the preferences. Concerning the policy problem, the issues were in general agreed upon:

there are many elements in the structural arrangements that push the universities and colleges in the same direction. We have nothing pushing them in different directions, there is steering based on the same goals and results, there are the same parameters and the same funding system, there is nothing pushing them to be different. (ministry)

Similar sentiments came also from informants from the sector. The aims of more system differentiation and division of labour, as well as strengthened profiles of institutions, were viewed as uncontroversial. The ministry also flagged a third aim, of strategic backing to institutional leadership, which created some ambiguity about the overall intentions. Thus, while there is agreement on the key problems within the system, the specific aims of introducing contracts were less clear.

Another source of ambiguity was the relationship to other steering instruments. While there was general agreement that complexity in steering mechanisms was an issue, there was uncertainty on how the contracts would enter this landscape. Those more sceptical would thus view the contracts as yet another set of goals in an already (too) complex and multifaceted governance landscape:

there has been a tendency after the Quality Reform that the main performance-based steering apparatus has been supplemented with more and more indicators, layer upon layer with different steering mechanisms and instruments. And this has been a rather unfortunate development. (HE sector)

This acknowledgement of uncertainty over the place of the contracts in the overall steering system is not only noted by sector representatives, also representatives from the ministry come with similar sentiments:

the steering system is too complex, there are layers upon layers with goals, there are too many indicators (…) We are painfully clear over the fact that we have goals from the law, we have sector level goals, we have institutional level goals, then we have now goals in the contracts … (…) the complexity of this all is just getting too much. (ministry)

Thus, while the open nature invited collaboration and co-creation, it also meant that the contracts obtained rather different forms and functions. The ministry had explicitly restrained itself from trying to standardise the process: ‘we really had to set this aim aside as we saw that institutions were concerned with various things. In the beginning, we were maybe more concerned about the homogeneity among contracts, but afterwards we have become more relaxed’ (ministry). The issue with this is that rather than shared understanding, a range of different ambiguities emerged: the function of the contracts (were they supposed to be instruments to steer the sector by, or for the institutional leaders to steer their institutions); and the kind of goals included (strengthen profiling or political steering, where one would require building on strengths, the other would imply focusing on shortcomings). Goal specification also represented a dilemma – the idea was to introduce goals that would not be typical quantitative indicators, there was also a concern about measurability and clarity of how to document goal attainment, consequently pushing the institutions towards more quantitative indicators.

Yet, the reasonably high degrees of trust suggest that there was also an underlying idea of developing better steering practices: ‘It feels like an attempt to say that we want to be a little closer to you, not to control, but to work together to develop important areas’ (HE sector). While a few of the sector representatives see the possibility of contracts opening up for more political steering, this was not considered a major issue at the time, suggesting some degree of alignment in norms and values of what good steering entails.

Overall, while Ansell and Gash (2008) place shared understanding as something that emerges from commitment, in the data here the two are much more intertwined, and commitment is also constructed through and strengthened by shared understanding, and vice versa.

Intermediate outcomes. While it is difficult to identify specific small wins already in the development process, the process did have a clear output in the form of contracts.

What kind of instrument did the contracts become?

A prime indicator of how the collaboration shaped the contracts is in the kind of instrument the conctracts became. The initial idea of contracts was them being a funding instrument, evident in the first policy proposal, the way in which the initial debates were framed, and also how the sector interpreted the signals. This had a strong influence on the kind of goals institutions considered. However, the contracts did over time become much more closely associated with the steering dialogue. Most of the informants involved saw their role as focusing, sharpening, and formalising the steering dialogue and elements that had previously been more informal. While literature on the use of performance contracts frequently discusses them in relation to funding (Jongbloed et al. 2018), the case here seems to suggest that in dialogue-oriented settings such instruments can also function without specific monetary incentives. Public higher education institutions in Norway are state owned and they receive most of their funding from public sources. Thus, the rather deeply internalised shadow of hierarchy is present. While most sector representatives viewed the contracts as dialogue instruments, this was not uniform and local implementation processes varied substantially (Elken & Borlaug 2020). Whether the contracts led to more differentiation is difficult to assess, as a sector-wide structural reform took place concurrently. This meant a drastic reduction in the number of institutions, and new kinds of system dynamics.

The data collection for this article was conducted in spring 2020. After data collection in this article, the system in Norway has undergone further significant changes. In 2021, a new White Paper was launched, outlining a reform of the whole governance structure (Meld. St. 19 2020–2021) to contribute to simplification and differentiation. The suggestion in the White Paper is to simplify the overall structure of goals and indicators, and that development contracts would receive a more central place in the overall system steering. It was announced that there would not be any funding attached to the contracts to fine-tune their function as a dialogue instrument and in this manner contribute to more differentiated steering.

In parallel, another debate had been ongoing about a new indicator for the funding system. In spring of 2020, the opposition obtained majority to decide that employability/relevance should be included into the funding system as a separate indicator. As the government was not behind this initiative, their solution to manage this was to establish a new expert committee to again review the whole funding system. Their proposal was put forward in spring 2022 and suggested radical changes – adding funding to contracts and removing all the performance-based indicators, aside two that focus on credit point and PhD degree production. In the White Paper by the ministry in spring 2023, this simplification was proposed, now pending approval from the parliament. The sector is now undergoing other rapid change processes and there is considerable uncertainty concerning future developments. The dialogue and consensus orientation in Norwegian higher education policy and governance can no longer be taken for granted.

Concluding reflections

In this article, the policy development of contracts – ‘development agreements’ – as policy instruments were analysed in the context of Norwegian higher education. Looking into the development of the contracts, the process was highly collaborative and open. As the expectation of possible funding involved was in the background, this also set a specific context for the process. Nevertheless, the instrument obtained a different form and became a dialogue rather than funding instrument. The rather specific contextual factors of a publicly funded higher education sector may have an impact here.

The case here also pointed at two inherent dilemmas: inclusion vs necessity to keep the discussions manageable; and necessity of openness but openness also creating uncertainty.

First, the analysis shows that while ordering the pilot in three stages had its benefits in terms of interaction and dialogue, there were challenges with exclusiveness/inclusiveness. The arena was not exclusive in that all institutions were invited to participate in one of the three rounds. Their capacity to set goals for their own contract was formally the same, but the discussions concerning the nature and purpose of the instrument seemed to become reduced over time. Thus, collaborative approaches like this include an inherent dilemma. The small number of participants in each round was seen as vital to assure good dialogue, but it also led to uneven opportunity to engage in the early discussions about the nature of the instrument. Collaborative governance thus has an issue of scaling – dialogue may necessitate to limit the number of participants, but sequencing or grouping is also problematic. This suggests that such governance approaches are either dependent on having a small number of relevant entities, or bodies that can in a meaningful manner represent the variety of different entities involved.

Second, the very open approach paradoxically both secured the process and created challenges. Openness was essential for facilitating involvement and commitment. Yet, it created uncertainty about what the purpose of the contracts were, and what the long-term outlook would be like. As such, the case illustrates that the ground rules established should not only concern the structure of the collaborative process but should also set a frame of reference for the discussions. Too open discussions can appear uncommitted and weaken opportunities to develop shared understanding, and in this manner weaken collaboration. This is the case despite reasonably high degree of trust within the system – even in a context of high trust there is dependency on clarity of expectations and ambitions.

What these two dilemmas particularly point at, is that collaborative approaches to policy development are not simple. Stronger involvement of the sector in development of the instruments that are there to govern them may seem appealing, but does not solve issues of power imbalance or the complexities of reaching shared understanding when institutional realities vary substantially. Most importantly, the case illustrates that not only does previous history of collaboration and conflict matter, the wider established governance regime matters – taken-for-granted assumptions of acceptable governing and acceptable use of governing instruments. In a high trust context, minor conflicts may therefore also not be an issue for the process. The study thus reiterates the focus on trust that some of the literature on collaborative governance. More importantly, the analysis also shows that instruments are not developed in a vacuum, they also interact with other previous, concurrent, and future processes. This may also provide enhanced complexity, as introducing networked and horizontal elements can also mean increased complexity when hierarchical elements are not being reduced (Lægreid and Rykkja 2014). Overall, while collaborative approaches can provide ways to strengthen legitimacy of governing, they also include dilemmas and challenges and therefore require careful design.

Acknowledgements

The author would like to thank Siri B. Borlaug who had an important role in collecting the data and initial discussions of the process, as well as two anonymous reviewers for useful comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The empirical data were collected as a part of an independent external evaluation of the pilot, funded by the Ministry of Education and Research.

Notes

1 Then NSD. SIKT was established in 2022 after merging several agencies.