Metagoverning social work knowledge structures

ABSTRACT

The increasing fragmentation and the rising number of organizations in society create major challenges when implementing knowledge structures. In this article, we propose that the concept of metagovernance is useful when addressing these challenges. The enhancement of knowledge structures in the Swedish social services is our study object. Networks at national, regional and local levels have been organized and facilitated by national resources and agreements since 2003. The shifting of participants, switching assignments and ambiguity about the networks’ purposes created difficult conditions for successful implementation. In this paper, we argue that fragile top-down government and shifting local network organization provide limited opportunities to ensure that new knowledge structures will have an impact on social work practice. Research is based upon monitoring a national program that started in 2003, affecting five universities and their surrounding regions and municipalities. The article is also based upon observations and interviews among national and regional actors. Results show that national agencies are hands-off, but still in control by setting up legal and discursive frameworks for those networks. It is argued that this creates new challenges for participating organizations and professionals in social work.

KEYWORDS:

• Metagovernance
• evidence-based practice
• networks
• government
• social work

Introduction

The execution of policy initiatives aiming to build knowledge structures and forming practice in social work have shown to be intrinsic. Proposals have been extensive; partnerships between academia and social work (Bledsoe-Mansori et al. 2013), increased service-user influence in knowledge development (Davies and Gray 2017), strengthened research in practice (Uggerhøj 2012), or increased influence for scientific knowledge production through international social science networks such as Campbell Collaboration together with reinforced efforts to implement research results (Palinkas and Soydan 2011).

Especially the idea that decisions should be based on the best evidence possible has become a megatrend, but not without challenges when implementing the idea (Boaz et al. 2019). The specific activities of knowledge production such as sophisticated evaluations need to be organized in order to support policymaking (Head 2016; Gambrill 2019). Ideally top-down recommendations and national guidelines should result in professional development and support interventions for the good of the clients. Such a linear, rational way of developing knowledge structures has severe limitations. One is the varying perceptions of what should be counted as reliable knowledge. Another – and one this article will address – is that contemporary implementation of policy is a complex process involving a multitude of actors at different levels. Never so well-grounded recommendations might fail when they reach practice regardless of their scientific base and strong political support.

In this article, we apply governance theory in order to analyse those problems. The realization of politics through top-down steering is out – networks are in. This has been suggested by many scholars of governance during the last 30 years (Barbeira, Franco, and Haase 2012; Jessop 2011; Hartley, Sørensen, and Torfing 2013; Huxham et al. 2000; Klijn and Koppenjan 2012; Lynn, Heinrich, and Hill 2001; Salamon 2002; Pierre and Peters 2005; Provan and Kenis 2008; Sørensen and Torfing 2007). In a move from government to governance, old regimes in bureaucracies now exist together with new flexible forms of negotiated governance in an increasingly fragmented society. When traditional hierarchal structures fail, politics and their agencies have to be more interactive by producing public policy through networks.

Within this broad literature, the related concepts of ‘metagovernance’ (Sørensen and Torfing 2007), ‘network governance’ (Provan and Kenis 2008), ‘new governance’ (Salamon 2002), ‘collaborative governance’ (Bingham 2011) and ‘new public governance’ (Klijn and Koppenjan 2012) seek to explain how the state is still capable of governing, but through other means and tools than before, such as coordination of partnerships and self-regulating actors. In particular, metagovernance tries to capture the attempts by influential political actors to govern self-regulating governance networks by setting limits and frames in the form of policy, resources, rules or discourse (Sørensen and Torfing 2007; Jessop 2011). Metagovernance is thus the ‘governance of governance’ (Kooiman and Jentoft 2009) and is used in this article to investigate how the state metagoverns self-regulating and self-governing networks at a distance in order to enhance social work knowledge structures.

According to Møller and colleagues (Møller, Elvbakken, and Foss Hansen 2019) Sweden is at the forefront as a country where national agencies strategically work to provide relevant knowledge for policy and practice. A national organization was launched in 1992 with the purpose of implementing knowledge-based social work along with methods to investigate and evaluate clients’ needs (Sundell et al. 2010). Different structures have since then been built up at the national level, in regions and in municipalities, in order to support the use of scientific knowledge. They are, for example, education, regional networks, guidelines and partnerships with universities that aim to produce and organize social work knowledge. Experiences from such various implementation processes within social services during a 15-year period in Sweden form the basis for this article.

Our research question is as follows: Which strategies of metagovernance can be identified when developing knowledge structures in social work? Investigating such strategies should contribute with new insights about how social work practice is governed and help us problematize knowledge management among actors with complex mutual dependencies.¹

The case of Sweden is built on two previously independent studies. They demonstrate a move from the first experimenting steps supporting the building of research-practice platforms (study 1) to the mobilization of several actors at different levels through negotiations and national and regional agreements (study 2). Data is drawn from the authors’ previously published studies of the implementation of knowledge-based social work. They are presented in section 4.2. Case study design is especially appropriate when investigating complex processes in their natural contexts (George and Bennett 2005; Yin 2003). Study one concerns a national program (KUBAS) where five universities and surrounding municipalities cooperated. They received generous grants and experimented with diverse forms of knowledge structures in the social services. The article uses data from an evaluation of one of those local projects (Sociorama) carried out 10 years after the official ending. In study two, the article uses studies at the national level where annual contracts between the state and the national actor SALAR have aimed to stimulate a more systematic development of evidence-based knowledge in public social welfare. SALAR is a significant organization that represents all Swedish municipalities and regions (www.skr.se). The studies mark the beginning and the prime of knowledge building ambitions in Sweden and represent two different forms of metagovernance.

We proceed by presenting arguments for a metagovernance perspective and how we aim to use metagovernance as an analytical tool to understand the initiatives for enhancing social work knowledge structures. Next, we describe our two studies, methods and data, and continue with an analysis of each one. Finally, we discuss our results.

A metagovernance perspective on the implementation of knowledge structures

The national implementation of knowledge structures in Sweden has been characterized as a guideline model (Bergmark, Bergmark, and Lundström 2011). This model builds on top-down steering of scientific-based knowledge by national authorities and requires compliance with knowledge of the best available interventions and assessments based on systematic literature reviews of published research.² It represents external and separate control over local decision makers, in which knowledge is formed in a scientific logic. This interest in science-based monitoring and decision systems has increased as the targeting and delivery systems have become increasingly common in the public sector. Hopes are raised that such systems will provide better decision support, making it possible to compare with other activities to reduce errors and contribute to increased transparency and democracy.

Welfare ministries and county councils have implemented a number of ingredients to promote this development. Each region is developing support structures in the social services and parts of the health care system with the goal of implementing best available knowledge in social services. The formation of various types of networks is assumed to be essential (Kommunförbundet Skåne. 2013) and SBU claims that they collaborate with extensive networks in Sweden as well as internationally in order to implement their findings and guidelines (www.sbu.se).³ A council of nine authorities was founded in 2015 with the mission of synchronizing issues of importance to the dissemination of quality-assured methods to be used effectively in health care and the social services.

The guideline model resembles the government type in the state’s movement in recent decades ‘from government to governance’ (Pierre and Peters 2000). Conceptualizing the implementation of knowledge structures as a guideline model might however understate significant self-regulating horizontal network activities and overstate the state’s ability to govern vertically by rules, not least due to the autonomy of local government. Instead, metagovernance emphasizes the control of self-organizing networks and the prescriptive regulation of autonomous networks (Haveri et al. 2009). This may involve framing of policies in existing governance networks, by influencing the allocation of resources and by intentionally designing the network (Sørensen and Torfing 2007). Such a perspective can problematize the governance network as a solution for knowledge management problems. Knowledge may take on a promotional and learning role for the members of these networks, although they may also try to control the external actions, such as the implementation of national guidelines. The two ideal types of government and metagovernance are described below (Table 1).

Table 1. From government to metagovernance.

	Government	Metagovernance
Organization	Bureaucracy, vertical	Network, vertical and horizontal
Primary actor relationship	Politics-administration	Influential political actors govern self-regulating governance networks
Role of state	Governing through rules and procedures and the production of national guidelines	Metagoverning through framing, institutional design, network facilitation and participation
Procedure	Following rules and guidelines	Cooperation, dialogue, negotiation
Rationale for legitimacy	Accountability, equality, input legitimacy	Coordination, problem solving, output legitimacy
Hanberger (2009, 4); Jessop (2011); Hertting and Vedung (2009, 57); Kooiman and Jentoft (2009); Pierre and Peters (2000, 14); Sørensen and Löfgren (2007); Sørensen and Torfing (2017)

Metagovernance as our analytical tool

In this article, we use the concept of metagovernance analytically, and not normatively. Even though it is a method of intervention for a metagovernor, our interest is to depict the state’s strategies for metagovernance. We define metagovernance as the governance of governance networks, i.e. the ‘horizontal institutionalization of the interaction of interdependent but operationally autonomous actors who collaborate in a shared effort to define and create public value through a process of regulated self-regulation’ (Sørensen and Torfing 2017, 829). Naturally, metagovernance shares characteristics with the concept of governance (cf. Hertting and Vedung 2009, 2012). First, by bringing actors with different perspectives, competencies and resources together in networks, the substantial and horizontal complexity of social problems can be explored from many viewpoints and solutions created (cf. wicked problems, Bingham 2011, 386). Second, mutual dependency causes the need for horizontal cooperation and self-organization (Klijn and Koppenjan 2012, 591). Third, the ideal network is open, flat, informal and temporary. This great flexibility can be perceived as a problem, counteracted if the networks organize a stable setting with long-term contracts. Another issue is the obscuring of the decision-making and selection of participants that are not only elected officials but also private or non-profit actors.

The fourth characteristic sets metagovernance apart from governance. It is the ability to govern the network as a metagovernor, i.e. ‘an actor, or a group of actors, who aims to initiate, support, and guide collaboration in governance networks to ensure that they contribute to the production of public value’ (Sørensen and Torfing 2017, 830). By distinguishing between the strength of the intervention exercised by this metagovernor and whether metagovernance is performed at a distance, or through close interaction, four basic strategies of metagovernance can be depicted (Sørensen and Torfing 2017, 829). Political, discursive and financial framing gives the network autonomy through overarching political goals, narratives or financial frames at a distance. Also, at a distance, institutional design means designing formal and informal arenas for a group of actors and defining the basic rules of engagement. This constraint of freedom of action can be achieved through explicit rules, procedures and structures of incentives, or by creating frames and narratives, which affect conceptions, interests and common identities in the network. An example is guidelines and preferred methods that postulate what knowledge to produce and disseminate aiming to influence the content of the networks (Sørensen and Löfgren 2007, 12). Apart from these distanced strategies, two more interactive strategies are possible. Network facilitation can help build trust and solve and avoid conflict by enhancing the network’s ability to define common goals, to coordinate their actions, and to support mutual learning processes. Lastly, the network participator actively participates in discussions and negotiations to gain influence on and affect the shared goals and strategies of the network (Sørensen and Löfgren 2007; Sørensen and Torfing 2017). An example is the way in which officials from the government and SALAR participate in networks and in conferences together with local and regional developers of knowledge structures.

Sørensen and Torfing (2017) argue that the four strategies should be viewed as complementary rather than as mutually exclusive. The former two ‘hands-off’ strategies are used at a distance, and could benefit from elements of the latter two more ‘hands-on’ strategies, and vice versa. Also, metagovernors should take care to balance the risk of overregulation which undermines the autonomy of the network, against the risk of underregulation, which leaves the network free from any overall direction (Sørensen and Löfgren 2007, 20). The success of networks as a governance strategy is thus based on the metagovernor’s ability to find the right level of regulation and interference (Sørensen and Torfing 2017). This underlines that the legitimacy of metagovernance is not built on deliberation or representation of stakeholders. The informality of networks in metagovernance might obscure decision-making and participation of non-elected or non-public actors. Instead, legitimacy is ensured by finding and presenting effective solutions for complex social problems. In the case of social work, metagovernance can be a way to engage in dialogue with civil society and non-profit organizations. However, participation and influence are dependent on the strategy the metagovernor decides to use.

In all, the characteristics and four strategies of metagovernance are ways to understand how metagovernors act and will be used to analyse how knowledge structures in social work are developed.

Case and data

The first study is an early example of sporadic financial support to self-regulating networks, while the second is a contemporary example of a nationwide implementation of knowledge structures. They represent two critical chapters of the Swedish state’s enhancement of social work knowledge structures while being connected to the international discussion of enhancing evidence through platforms and partnerships (Bellamy et al. 2008; Palinkas and Soydan 2012). Our use of two studies is not comparative but serves to illustrate strategies of metagovernance for social work knowledge structures. Next, we describe the studies, data and analysis.

Two steps in the development of social work knowledge structures

Sociorama started in 2001, when the NBHW was instructed by the Swedish government to create knowledge development in the social services through regional cooperation between research, practice and education (the KUBAS-project). Until then, knowledge support had been initiated, produced and disseminated by the NBHW. The new aim was to financially support the integration of local social service organizations with social work research and education. Universities and municipalities applied for funding (NBHW 2005). Apart from reviewing and financing the projects, the NBHW was not involved locally. The Sociorama project in the Kronoberg region was one of five KUBAS projects (knowledge-based social services) which were run at university sites. The diverse KUBAS-projects had great resemblance to other partnerships at the time aiming to develop cooperation between a university and social work practice, like the San Francisco Bay Area partnership (Anthony and Austin 2008; Austin et al. 1999) and the Matilda Wrede Institute, a Finnish research institute (Julkunen 2011). Sociorama aimed at creating structures for greater integration between research, practice, and education in the field of social services. It lasted from 2003 to 2009, most actively from 2003 to 2005. One component was shared staff positions between a university and the social services of a municipality (henceforth ‘shared positions’). Others were collaborating with user organizations, research circles, joint education and a party-based steering committee. Sociorama was a trial, a project that created ‘cross-border knowledge development’ according to the head researcher (Salonen 2005, 6). Developing and implementing new working methods was not on the agenda. When the government’s KUBAS money ended in 2005, Sociorama was integrated in an existing regional R&D.

National agreement is another chapter in the story of Swedish knowledge-based social work. According to the commission ‘Evidence-based Practice in Social Services – For the Benefit of the Client’ (SOU 2008, 18) the knowledge base in the social services was undeveloped and interventions were to be scientifically supported. Representatives of central and regional governments signed a platform agreement in 2009 to strengthen the social services’ ability to use the best available knowledge. Governmental agencies together with SALAR then reached six annual working agreements to support knowledge development in the social services (Överenskommelse 2009–2015). This included forming support structures at the regional level through county councils. Incentives were introduced under the designation supportive structures. Managers were hired, and several projects started. The government asked regions to formulate joint declarations of intent. Regions that sent such declarations to SALAR were granted funds. The two cases are summarized below (Table 2).

Table 2. Characteristics.

	Sociorama	National agreements
Active years	2003–2009	2009–2016
Metagovernor	NBHW	Ministry of Health and Social Affairs, SALAR
Purpose of network	Create a structure for the integration of social work practice, education and research	Build social work knowledge structures by supporting the national implementation of EBP
Governance network actors	Municipality, university, region, non-profit organization	SALAR, regions/counties, municipalities (through regional platform leaders)

Informants and data collection

The first study is based on an evaluation by first and second author 10 years after the start of Sociorama (Denvall, Lernå, and Nordesjö 2014). Data was collected through telephone and individual interviews, focus groups and a survey. Telephone interviews were conducted with 26 social services managers in 20 municipalities. 11 individual interviews took place with previous coordinators of Sociorama and 8 with people working at the university, the larger municipalities, a non-profit organization and officials from two regions. 20 persons with shared positions in Sociorama, who today were working in different social work-related areas, answered a survey. Finally, three focus groups were carried out with 4–6 informants where most (11/16) had already been interviewed or participated in the survey. In total 61 informants participated. Informants were selected due to their participation or decision-making function in Sociorama or today’s local knowledge-based social work. When making contact, most were easily recruited to the evaluation since they had many thoughts and memories of the project. The interviews were semi-structured where questions would follow up the opinions of the project 10 years after, the municipalities’ present knowledge-based social services, and the cooperation between the social services and the university. For example, themes related to Sociorama were the importance of Sociorama today and methods from Sociorama you use today. The survey’s questions were predominantly open-ended and identical to the interview questions in order to compare answers. Telephone and individual interviews took 20–45 minutes and focus groups 90 minutes. They were all recorded and transcribed verbatim by the authors. Data was compiled in brief summaries and interpreted, validated and corrected at three seminars with people who had worked in the project, managers from the social services and officials at the university. Governance-theory were used to analyse data.

The second study is based on empirical work related to national agreements for evidence-based practice (EBP) carried out by the third author between 2009 and 2016. In the first published article (Denvall and Johansson 2012), the implementation of EBP referred to the Swedish government’s take on how to create knowledge-based structures through evidence. The first of three steps were the collection of documents and reports such as policy proposals and evaluations by national government commissions related to the agreements. The author also participated in workshops, conferences, seminars and meetings aimed at anchoring and spreading the national idea of EBP, where actors from national, regional and local organizations participated. Program-theory were then used to analyse data. In step 2, 30 managers and experts on EBP and its Swedish implementation from the national and regional government were interviewed by third author 2010–2015 to learn more about the implementation (Johansson, Vedung, and Denvall 2015). Also, a survey distributed to 317 professionals with a response rate of nearly 60%, was conducted in 2015. The survey included seven interviews and focused on the professionals’ views and understanding of EBP. In a third step, a small participant observation of a local attempt at network governance was conducted in 2014–2015. This observation study included three observations, each about six hours long, during the start of this network (Johansson and Fogelgren 2016, 2019; Johansson 2019). Translation theory and institutional theory were then in use to analyse data.

To summarize, this article is based on work over a series of 8 years conducted by six researchers. The studies were published on seven occasions during the period 2012–2019 as seminar-treated reports at two universities (Denvall, Lernå, and Nordesjö 2014; Johansson and Fogelgren 2016), as a chapter in an edited book (Johansson 2013) and as four peer-reviewed articles in journals in social work and in social science (Denvall and Johansson 2012: Johansson, Vedung, and Denvall 2015; Johansson and Fogelgren 2019; Johansson 2019).

Data analysis

For the purpose of this article, we reinvestigated the data from the perspective of metagovernance as described in a previous section. Depicting strategies of metagovernance in each study was done by using the four characteristics of metagovernance as analytical themes in a theory-driven analysis (cf. Braun and Clarke 2012). Data from the two studies was thus read and structured according to characteristics of metagovernance. For example, in the first study, the characteristic and theme of substantial and horizontal complexity concerns to systematic knowledge development and the integration of practical and theoretical knowledge in social work practice, while it in the second study concerns the Swedish process of implementation of EBP.

Analysis

The strategies for enhancing knowledge structures in social work are analysed below in terms of metagovernance. The analysis is summarized in Table 3 according to the characteristics of metagovernance.

Table 3. Summary of analysis.

	Sociorama	National agreements
Achieved knowledge-based structures	Reorganizations and lack of funding changed the commitment and incentives for network actors The network faded away, and few ideas and no structures have survived ten years later NBHW concludes that municipalities and universities cannot finance these structures long-term, and suggests several state-funded institutes of social research	Some regions have built up support structures for knowledge production, although the content that is to be implemented remains unclear Improved conditions for using evidence, although still a long way to a more systematic development of knowledge Unknown impact in local social services
1. Substantial and horizontal complexity of social problems	Yes – Problems of systematic knowledge development and integration of practical and theoretical knowledge addressed by actors with different knowledge and from different geographical and institutional areas.	Yes – The problem of the implementation of EBP. Actors with shifting knowledge perspectives and areas of responsibility.
2. Mutual dependency between network actors	Yes – Mutual dependency weakened over time due to limited resources and lack of financial agreements.	No – As a state-driven project, network actors were not mutually dependent and lacked incentives for self-organization.
3. Open, flat, informal and temporary networks	Yes – Sociorama was open, flat and informal, making it sensitive to changes in the actors’ and the metagovernor’s interests and needs	No – The metagovernors controlled the interpretation of objectives and means and the participation of actors.
4. Strategy of metagovernance	Political, discursive and financial framing: Local practice and research, not the state, should produce social work knowledge. NBHW coordinates and supports financially	Institutional design: Governmental agencies and SALAR reach annual agreements to support knowledge development in the social services. SALAR forms support structures at the regional level through county councils. Regions must accept rules and agreements to be a networking partner
4a. Strength of strategy	Autonomy for Sociorama actors within discursive frames Mutual dependency between state and network and within network	Autonomy for network actors within specified rules
4b. Challenge of strategy	Metagovernor is passive and far away, network vulnerable to changes in financial framing	Emphasis on rules and procedures resulted in unclear content Lack of mutual dependency weakens ties among actors

Sociorama – political, discursive and financial framing of local knowledge

The Sociorama project attempted to create a structure for the integration of social work practice in the Kronoberg region and education and research at the Department of Social Work at the university. It was regarded as important to build relationships and get to know each other’s situation and find ways to collaborate. Among several collaborative activities were shared positions between the social services and the university. Those who themselves had shared positions emphasize personal gains such as improved collaboration, knowledge and their own careers. They could bring practical knowledge into research, and research knowledge into practice. Additionally, informants found research circles successful where researchers and practitioners met and discussed a specific topic. Also favoured was BIKVA evaluation, where evaluators start from service users’ conceptions to create criteria for the assessment and development of social work practice (cf. Krogstrup 2017). Despite such positive experiences, informants claim that few of the tested ideas have survived to the present day.

From a metagovernance point of view, the Sociorama project was an answer to problems of both substantial and horizontal complexity. Problems of systematic knowledge development and integration of practical and theoretical knowledge could not be addressed by one actor alone. They would have to be addressed with the help of actors with different knowledge such as practical and theoretical knowledge, and from various areas of the social services and within social work research (substantial complexity). Actors also represented shifting municipalities and regions, as well as areas within the social services, holding dissimilar mandates within the network (horizontal complexity).

There was also a mutual dependency between the state and the network’s actors, and within the network. The state wanted social work knowledge to be formulated locally, and the network was in need of resources and guidance to experiment with knowledge-based structures. However, the mutual dependency between network actors weakened when government resources ended in 2005. Without external money, it became more difficult to coordinate common resources and strategies. Initially, the main resource was participants’ working time, but this ‘invisible’ cost suddenly became visible, as Sociorama had to be financed through other means. Also, no financial agreements were made prior to the new financial conditions which could have stabilized the transition between autonomy and integration in a regional R&D unit. According to informants, Sociorama’s members lost their focus and commitment when implemented into a new organizational environment without independence.

Sociorama was relatively open, flat and informal, making it sensitive to changes in actors’ interests and needs. It went from being a KUBAS project with high autonomy, to being integrated into an existing regional platform as one among several welfare work areas. The university lacked commitment since it was concentrating on an upcoming fusion with a university college nearby. Municipalities turned to themselves where some saw Sociorama as a competitor to other development initiatives. As the years went by, leading municipal positions were replaced which made long-term commitment to suffer and resulted in much more blurred interests and incentives. After a while, informants say, it was unclear if the new project management was running Sociorama as before, or as a sub-unit of the regional R&D. The collective problem solving that earlier on had resulted in activities within Sociorama, gradually was intertwined in new conditions of authority, delegation and overarching goals. In short, unclear and unstable borders of the network were easily challenged when organizational conditions changed.

NBHW’s role was to act as a metagovernor in financing Sociorama within the KUBAS program until 2005. It is a form of political, discursive and financial framing where a government agency metagoverns from a distance without participation and facilitation, but through an accepted project plan describing allocated financial resources and possible actions within the discursive frames of Sociorama. As Sociorama was not fully funded by NBHW, metagovernance was rather distant and presumably less interventionist. Still, the metagovernance strongly facilitated the self-organization of network actors where mutually dependent actors partly funded the project themselves since they identified problems impossible to solve on their own. The financial aid was decisive.

In conclusion, Sociorama can be understood as political, discursive and financial framing that in many respects differs from today’s more interventionist ways of implementing knowledge structures. This metagovernance does not necessarily call for direct interaction between the metagovernor and the networks. Instead, a low level of intervention is a part of the bargain. For Sociorama, this meant a high level of autonomy for a network of mutually dependent actors. But the lack of resources combined with diffuse and open borders easily challenged the network and Sociorama slowly faded away. This resonates with an overarching conclusion from the NBHW (2005) regarding all KUBAS projects: municipalities and universities could not finance these structures in the long term.

National agreements – institutional design with shifting content

In our second study, several projects were introduced in the regions concerned under the designation supportive structures for an evidence-based practice for good quality in the field of social services. The first step by the government was when all regions were asked to draw up joint declarations of intent (Johansson 2013, 2019). All regions were granted funds to continue working on support structures if these joint declarations were sent in to SALAR. During this intervention, relations between the state and providers changed as new actors, methods and demands emerged as the local social services were supposed to implement and adapt novel decision-making tools and systems (Denvall and Johansson 2012; Lundström and Shanks 2013; Billsten et al. 2018). An example is the emergence of about fifty process (platform) leaders with the mission to initiate local change projects and support the implementation of EBP. They were organized as two national networks, one within children and youth investment and one within abuse and addiction care. Another was the implementation of national guidelines for abuse and addiction care and the support for the development of more long-term structures for the provision of knowledge, including training of social workers and the establishment of networks within and between counties (Billsten; 2019, 91). Since the start, research and development activities have been transformed and reorganized several times within regions, municipalities and universities.

Just as in the case of Sociorama, the national agreements tackled a complex issue – the nation-wide implementation of EBP. They mobilized a multitude of national and regional actors, researchers, officials, professionals and service-users, often with shifting knowledge perspectives (substantial complexity) and areas of responsibility (horizontal complexity). Several efforts were made within the framework of the annual agreements, but the outcome and efficiency and the long-term sustainability of these support structures have been challenged. By utilizing Ernesto Laclaus’s theory of empty signifier Gegner, Righard, and Denvall (2019) were able to show that the regional process leaders used evidence as a floating term that was filled with different kinds of content depending on the purpose. The process leaders described EBP as a model of governance, not in terms of knowledge. Other research shows less successful effects, loss effects (one of the ideas was that the work should be more beneficial, systematic and standardized), for users and for the professionals (Liljegren and Parding 2010; Martinell Barfoed 2014). The regional levels are more or less back in the same situation, regarding knowledge development that existed before this intervention (Johansson and Fogelgren 2016).

In terms of metagovernance, the lack of mutual dependency is striking. Since the implementation of EBP is a state-driven project, participating actors are not mutually dependent and are therefore lacking incentives for self-organization. This makes metagovernance interventionist. As we have seen, the implementation of EBP is more complicated than simply giving orders to the social services and social welfare practice in general to benefit the client through using the best knowledge available. It consists of organizing and directing citizens, professionals, organizations, institutions, and discourses to achieve certain political goals. The metagovernor strove to build flat and horizontal networks of equal partners outside the social services practice where consensus was supposed to be a central mechanism (Johansson 2013). It turned out, however, that SALAR and NBHW had the power over the interpretation of objectives and means and that other stakeholders such as researchers, officials, professionals and service-users had less real influence on the future development of knowledge in the social services. For example, national networks organized by SALAR were challenged by ambiguity about how to interpret the concept of evidence, which created unclear assignments and responsibilities for local governments. It is well-known that implementation processes do not always follow a vertical organizational logic even if rules, procedures and arguments seem to be at hand. In this case street-level bureaucrats interpreted EBP in a way policymakers did not expect. This illustrates both shortcomings of hierarchical government and insufficient metagovernance.

The main strategy for the metagovernors SALAR and NBHW is thus institutional design. Like political, discursive and financial framing it is also used at a distance (Sørensen and Löfgren 2007, 13). However, the strategy institutional design is highly interventionist since its aim is not only to create space for autonomous networks, but also to influence the content of the self-regulation. This is certainly true in this case where agreements and support structures were thought to work as incentive steering and create collective points of identification and meaning under the umbrella term supportive structures for EBP.

In sum, the second study shows serious challenges in the implementation of social work knowledge structures. Although cooperation and networks on multiple levels were initiated through institutional design, the local networks were given troublesome tasks. The meaning of evidence and ways to achieve it were unclear and expected to be supplemented by negotiations locally that should be clarified and processed in the future. Agencies were under constant reorganization, shaping ambiguous surroundings with fragile networks. Still, the aim of building social work knowledge structures has strong support at the national level, which affects local agencies who are preparing for something they suppose will support efforts to improve their undertakings, although it is unclear how. Today, the formal initiative has come to an end and the Swedish Agency for Public Management argues in its final evaluation that the regions have built up support structures for knowledge production within networks of people (Statskontoret 2014, 8). They note that it has improved conditions for EBP. However, ‘It is still a long way to a more systematic development of knowledge’ (ibid., p. 7).

Discussion

We start the discussion by returning to the problems this article addresses and to its purpose. Then we summarize the results and, in more depth discuss two of those before we end with recommendations for practice, for national authorities and for research.

Ideas for using scientific knowledge for evidence-based policy and practice have had a huge impact on management and research. Marketization and the emergence of new actors have formed new kinds of public-private-civil partnerships, together shaping a wide range of organizations and thus affecting the possibility to implement policy. Organizing networks with multi-actor engagement across organizations represents a means of implementation and innovation (Hartley, Sørensen, and Torfing 2013). Research has demonstrated the difficulty of solving problems by control in hierarchical structures and shows that the application of standardized systems for knowledge management is determined in the local context (Barfoed and Jacobsson 2012; Boaz et al. 2019; Greenhalgh et al. 2004; Hjelte, Brännström, and Engström 2010). The development of knowledge structures when implementing EBP in Sweden corroborates this research.

Current research has introduced the concept of metagovernance which tries to capture how the state and its authorities still govern but in other forms and by other means than before. In this article, we asked which strategies of metagovernance could be identified when developing knowledge structures in social work. With the help of the presented studies, we have demonstrated how political, discursive and financial framing (Sociorama) and institutional design (national agreements) were vital yet deployable strategies within the concept of metagovernance. Sociorama was an autonomous, temporarily stable, and nationally funded network within the frame of knowledge-based social work, but soon lost momentum and support when resources and other preconditions changed. In the case of national agreements, actors on a national level put relatively sharp limits to what networks were to work with. This has created activity at the desired regional and local levels, but lack of sufficient resources and clarity about the contents of the networks, as well as input from science, explain its lack of outcome that might support social work practice.

We want to highlight two of the results and to pose a question. First, we suggest that the concept of metagovernance may deepen our understanding of the relationships between actors who are involved in building knowledge structures in social work. It gives us the possibility to understand incentives and agency beyond rule following and competition. Challenges in creating knowledge structures in our two examples can thus be understood in terms of unclear framings, network participants’ changing interests and lack of participation in the network’s problem definition. For example, through institutional design, national actors can set relatively clear limits to the networks and their missions. The metagovernor may affect the network’s self-regulation through strategic construction, the creation of incentives and the internalization of a clear collective identity. Similar to political, discursive and financial framing, it is a remote strategy without directly involving the network. These insights have been valuable to understand knowledge structures in the Swedish cases by focusing not only on the (meta)governors but also on the networks at regional and municipal levels that are left with various degrees and forms of governance. We thus agree with studies that has focused more on the actors and relations between actors in the governance of EBP, than on its methods and implementation (Elvbakken and Hansen 2019).

Second, it is possible to understand the strategies through which social work is not metagoverned, i.e. through network facilitation and participation. Our analysis suggests that the government has not been active in facilitating for actors to solve their network challenges. Instead, they rely on framing and agreements on a national level. This brings us to our question: how far does the more distant metagovernance strategies reach? The national government was distant and seemed to be hoping for the best. The challenges in supporting the emergence of strong knowledge structures in social work can then be a misjudgement of the network’s ability to create results (theoretical error) or poor implementation (implementation error). Although bottom-up implementation through networks may seem like an adequate strategy in dealing with the complexity of wicked problems, an actor needs to support and create necessary conditions for the network. Participants shift, switching assignments and uncertainty about the networks’ objectives and how to define their tasks and participants, all combine to create demanding conditions for successful implementation. Thus, weak top-down control and a weak horizontal-up network organization provide limited opportunities to ensure that scientific knowledge will support policymaking. This central problem is in line with an essential idea in metagovernance theory and research – to balance the risk of overregulation against the risk of underregulation. In this case, it may very well mean that, as Sørensen and Torfing (2017) argue, interventions (such as the identified hands-off strategies), might benefit from the complementarity of other strategies such as facilitation and participation in relation to local networks’ needs.

One recommendation for practice in social work is that further advance of knowledge structures should include development of the local social services’ organizations and their leadership. This research only deals with metagovernance within the building of knowledge structures in a developed welfare state. However, with increasing tension between different groups and erosive welfare systems the state’s limited ability to govern is worrying and should be subject to further knowledge development. A second recommendation based on this study will be that the national authorities need to develop their ability in facilitating and participating in metagovernance. This could be in the form of facilitating networks or actively participating in them. Research has previously demonstrated shortcomings of central authorities when attempting implementation and their dependency on information and sermon (Jacobsson and Meeuwisse 2018). This study has shown that the implementation is counteracted by changing and insecure networks. The state has limited power to change the practice at a distance from the top with temporary resources. Metagovernance needs to be long-term and in the form of network facilitation and active network participation which should lead to the state and its authorities also need to develop their ability to support for local metagovernance. This could be in the form of facilitating local networks or actively participating in them. At last, we find it surprising that meta-governance has become so limited in scope as an analytical tool in social work research. In our opinion, it contributes to a problematizing perspective on governance and provides practical guidance for actors when designing new forms of governance. This is especially relevant in the area of social work where the ‘who’ might be one or several metagovernors on several different levels (cf. Gjaltema, Biesbroek, and Termeer 2020). A final recommendation for international research is to a greater extent publish in journals related to social work and to the editors of these journals to invite to topic-issues on the current forms of governance.

To conclude:

The governance of social work in municipalities has increased in scope, complexity and level of detail in recent decades. New governance forms have been added, in the form of, among other things, agreements, action plans and knowledge management. Government through national guidelines is being supplemented by the metagovernance of networks. National agencies are distant, but still in control by setting up legal and discursive frameworks for how those networks should function. This increasing metagovernance is also described as short-term, jerky and not adapted to local conditions. Such institutional design creates new challenges for participating organizations and professionals in social work. Social work research needs to go beyond humble criticism of hierarchical implementation problems and instead take an interest in forms of institutional design, readiness for change and, not least, actors’ participation in metagovernance.

Acknowledgments

Very sadly third author Kerstin Johansson passed away early Spring 2020. We would like to highlight her great contribution in the preparation of this article and the studies on which it is based.

Disclosure statement

We have no potential conflicts of interest with respect to this research, authorship and/or publication of this article.

Additional information

Funding

We received initial financial support for this research by Linnaeus University and Linköping University, however no funding for the publication of this article

Notes

1. The development of evidence in a Scandinavian context has been extensively described by Møller, Elvbakken, and Foss Hansen (2019) and in Sweden by Sundell et al. (2010). Challenges in this issue have been reported by for instance Avby, Nilsen, and Dahlgren (2014), Björk (2016) and Jacobsson and Meeuwisse (2018).

2. National authorities such as The Ministry of Social Affairs (Socialdepartementet), the National Board of Health and Welfare (NBHW, ‘Socialstyrelsen’) and the Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU) are regularly developing updated national guidelines for health and social care based on the best available knowledge. During 2019 for instance about mental illness and addiction.

3. http://www.sbu.se/en/about-sbu/publications-and-communication/, 17 February 2019.