How to frame the governance dimension of social innovation: theoretical considerations and empirical evidence

Abstract

Social innovation approach has been increasingly used by governments in delivery of public services, since the development of more and more complex societal challenges requires the establishment of new multi-actor implementation structures and arrangements. More recently, a call for a more robust analytical framework emerged in order to enable evaluation of the growing number and types of socially innovative practices implemented in different European contexts. This article takes up this challenge by assuming the governance perspective on social innovation, i.e. the establishment of new implementation arrangements in relations between the public and non-profit sectors. Drawing on the public governance literature, the article describes a three-step methodology with which to design and implement socially innovative oriented governance, and it illustrates an empirical application to the issue of refugee integration. The article argues that the proposed methodology is suitable both for assessing if and to what extent civil society organizations are actually involved in horizontal and cooperative relations with public actors when new implementation tasks are required, and for guiding scholars and practitioners in investigating what should be improved to achieve socially innovative governance within a public policy process.

Keywords:

• Social innovation
• New Public Governance
• co-creation
• multi-actor implementation arrangements
• refugee integration policy

1. Introduction

In the past decade, national and sub-national governments have had to deal with several environmental and societal challenges that has required the production of new public policies (Steinebach and Knill 2017; Casula 2022). These policies entail new implementation structures (Peters 2015), and the establishment of new multi-actor implementation arrangements (O’Toole 1986) that involve multiple public and private bodies (Thomann, Hupe, and Sager 2018; Steinebach 2022), and that require the consolidation of new interorganizational relations to improve the delivery of public services (O’Toole 2012).

In this changing context, since the turn of the new millennium, the concept of Social Innovation (henceforth ‘SI’) has been increasingly included among the ‘magic concepts’ (Pollitt and Hupe 2011) used by both scholars and policy-makers to frame and support solutions to deal with new societal problems and challenges (Broekema, Horlings, and Bulder 2021), usually with a focus on the sub-national level. Nevertheless, as with many ‘magic concepts’ in social science, a number of definitions exist, and they often compete with one another. A shared operational definition is also lacking. Thus, there is a risk that every new social measure may be labelled or self-defined by the proponent as SI. To remedy this drawback and make possible the measurement and evaluation of practices and policies characterized by new multi-actor implementation structures that strive to be socially innovative, scholars have called for a specific operationalization of SI. Achieving this goal would be of paramount importance both to increase the understanding of SI at the scientific level and to provide policy-making tools useful for making decisions in the face of complex societal challenges characterized by high uncertainty (Krlev, Bund, and Mildenberger 2014).

As a result, during the last decade a pioneering literature on SI metrics developed (Reeder et al. 2012; Krlev, Bund, and Mildenberger 2014; Bund et al. 2015). However, it displayed a set of deficiencies. Firstly, the identified SI metrics obtained scant consensus among scholars concerning their reliability and robustness of their preliminary findings. This was primarily the consequence of disagreement on the definition itself of SI as a concept (Mihci 2020, 357). Secondly, it is now widely acknowledged that SI is context-dependent, cross-sectorial, cross-disciplinary, and cross-geographical (Moulaert et al. 2005; Benneworth and Cunha 2015): adopting a one-size-fits-all approach may thus be neither feasible nor appropriate (Cunha and Benneworth 2020).

Against this background, the article intends to respond to the call by scholars for an improved definition of a conceptual framework in which to analyse social innovation initiatives and practices that require the establishment of new implementation arrangements to deal better with new societal challenges. However, instead of looking for specific indicators, and therefore for specific metrics related to the outputs (social innovative practices), the article focuses on the process of SI, i.e. on the procedure for designing a socially innovative governance system. It argues that the construction of social innovative governance is largely in itself the construction of SI. The emergence of new public problems and governance challenges, together with the decline of the state’s regulatory capacity (Lodge and Wegrich 2014), have in fact brought innovation and the necessary transformation of relations among state, market and civil society to the centre of attention.

Taking advantage of the political science literature on the different waves of public administration reforms, this article starts with the assumption that a precondition for non-ephemeral SI is the construction of socially innovative governance systems in which public and private actors, together with the beneficiaries of an intervention, enter into a close relationship of collaboration. This is a co-design and co-creation activity where both state and non-state actors are committed to coping with intractable and wicked problems related to social exclusion. Socially innovative governance is in fact particularly valuable in addressing ‘growing social challenges that neither government nor citizens have the necessary resources to solve on their own’ (Pestoff 2012, 1106).

Given these premises, as extensively explained in the next section, the aim of this article is threefold: (1) to improve the theoretical understanding of SI specifically related to the analysis and evaluation of what a socially innovative process is; (2) to use the identified framework in order to develop a specific methodology with which to design and implement a socially innovative governance system; (3) to apply this methodology empirically to the issue of refugee integration in order to test the strengths and possible drawbacks when the methodology is implemented.

We believe that this article can make both a theoretical and practical contribution. As regards the former, it provides scholars and policy analysts with a toolbox with which to assess whether new social practices claimed to be socially innovative are actually informed by a socially innovative governance arrangement. As regards the latter, it provides policy-makers, practitioners and third-sector organizations (henceforth ‘TSOs’) with viable solutions when they intend to promote social innovation practices inspired by innovative governance.

The article draws on the methodology designed and implemented within an ongoing European collaborative project dealing with refugee integration in society and the labour market. The project specifically aims to enhance coordination and virtuous governance among political institutions, TSOs and local communities (including refugees themselves) in designing, implementing and evaluating policies related to refugee integration at the local and transnational level.

The article is organised as follows. Section 2 presents the theoretical background, including the debate on measuring SI and a literature overview focusing on the changing operational paradigms in government and particularly the governance dimension of SI. Section 3 introduces the three-step method for designing social innovative practices, Section 4 describes an application of the framework to refugee integration policies. Section 5 concludes and discusses both how third-sector studies can benefit from a social innovation framework in governance arrangements, and the methodological implications that arise when studying the third sector with our approach.

2. Theoretical and analytical background

2.1. The debate on how to measure social innovation: an appraisal

One of the reasons for the difficulty of identifying a way to evaluate SI is the lack of a shared definition of SI. It is therefore convenient to start from SI as a contested concept, i.e. one that is internally complex and susceptible to different interpretations and changes over time (Ayob, Teasdale, and Fagan 2016). As illustrated by Ayob, Teasdale, and Fagan (2016), after its emergence in the early 1990s – sporadically, without an actual process of knowledge accumulation – between the mid and late 1990s publications focused mainly on the social impact of technological innovations (e.g. the vacuum cleaner, which produced greater sharing of tasks within the household). Since the beginning of the 2000s, while the technology-oriented literature has continued, a new trend has emerged. It focuses on SI as new forms of social relations and on how they can generate better social outcomes. Over the past fifteen years, an emphasis on SI as a process leading to societal change has emerged, and it now predominates. The focus on technological change, however, has not entirely disappeared, even though it has become a minor concern. The process dimension is one of the three dimensions of SI identified by Moulaert et al. (2005, 1976): it implies changes in social relations, especially regarding governance, that allow the satisfaction of social needs and increase the participation of deprived groups.¹ Since the last decade, SI has gained considerable attention both in academia and within policy circles, being considered an essential means to modernise welfare states. And it is precisely at this time that there has emerged the absence of an operationalization of the concept of SI which would allow it to be measured and therefore evaluated. Indeed, SI requires a mode of measurement distinct from the focus on technological-economic indicators like R&D expenditures in the private or public sector, or the number of patents (Reeder et al. 2012; Bund et al. 2015). However, the assessment of SI encounters some complicated issues: for instance, it is deeply embedded in the social fabric of a specific community; it is not a specific sector but cuts across different sectors and activities; it is changeable since it reacts to the specific context (Reeder et al. 2012, 11–12). Moreover, SI often does not produce a tangible output/product, being much more closely related to the development of a principle, an idea, or an intervention (Cunha and Benneworth 2020).

The European funded project TEPSIE (The Theoretical, Empirical and Policy Foundations for Building Social Innovation in Europe), which ran from 2012 to 2014, tried to bridge this gap by developing the blueprint for an indicator system to be used in the design of the EU strategy for SI. The proposed metrics assumed the macro level of analysis, i.e. the nation-state, partly because it allows for a cross-national comparative perspective on SI, due to the comparative and aggregate nature of macro-level indicators, and partly because it permits the application of measures of innovation by commercial and public sectors (Krlev, Bund, and Mildenberger 2014). The TEPSIE project is undoubtedly an important methodological contribution, and it is cited by every study dealing with the topic of measuring SI. One of its main merits is that it addressed both the need for further understanding of the concept of SI and the need for data to be available to decision makers in the policymaking process. Regarding the improved conceptual understanding of SI, TEPSIE identified three dimensions of analysis: (i) framework conditions enabling innovation, which comprise conditions at the ‘institutional’, ‘political’ and ‘social climate’ level; (ii) entrepreneurial activities, including the categories of ‘investment activities’, ‘start-up activities and death rates of the firms’ and ‘collaboration and networks’; (iii) output and societal outcomes, which concern many dimensions such as ‘education’, ‘health and care’, ‘employment’ ‘housing’, ‘social cohesion’, ‘political participation’ and ‘environment’ (Bund et al. 2015). As regards the project’s contribution in terms of data usable by policy-makers, it developed approximately 120 quantitative indicators within the three above-mentioned dimensions. However, precisely this main finding of the project (the high number of indicators proposed) is also one of its main limitations, because obtaining reliable results is extremely difficult with more than 100 indicators (Mihci 2020). Moreover, Mihci (2020, 349) underlines three further main limitations of the measurement procedure propounded by TEPSIE. Firstly, the methodology is overly dependent on traditional systems of innovation indicators, i.e. the technological innovation indicators used by the European Innovation Scoreboard. Secondly, data sets for some of the proposed indicators are not actually available, so that the success of the SI computational process is threatened. Thirdly, most of the indicators – like ‘educational attainment’ or ‘PISA results’ – have static rather than dynamic characteristics (i.e. they capture the situation in a particular moment), although the latter seem much more suitable for identifying the social outcomes of SI. The other methods used to measure social innovation (Bund et al. 2015; Castro-Spila, Luna, and Unceta 2016) that Mihci (2020) has mapped also have drawbacks, especially ones related to an excessively narrow scope that does not allow any extension to a supra-local level.

Overall, the attempts made so far do not seem satisfactory. Consequently, it seems necessary to change the approach that measures social innovation as a product using metrics, and instead narrow the focus to a conceptual framework in which to analyse the SI process and particularly its governance dimension.

2.2. In search of a conceptual framework for SI: bringing the state back in

Our emphasis on the governance dimension of SI is intended to reinstate the role that public institutions should play in the production of SI-oriented service provision. In fact, because part of the literature on SI claims the leading role of third-sector organizations in engendering social innovation practices, it tends to consider the intervention of public institutions – particularly at the national level – as an obstacle and a threat to the autonomy of civil society organizations (see Abad and Ezponda 2021 for a detailed reconstruction of this debate). Indeed, the relation between public institutions and SI is sometimes considered problematic (Lane et al. 2009), as is also apparent in the tension envisaged between social innovations that follow bottom-up, i.e. grass-roots based, trajectories and institutionalization: recognition and support of public authorities are sometimes not deemed useful for SI, since public actors are prone to bureaucratization and this could distort the very meaning of SI (Moulaert et al. 2005). According to this stream of literature, SI is a strategy with which to overcome the bureaucratic rigidities of the public sector, which is considered unsuitable for dealing with complex problems, also due to its risk aversion (Young Foundation 2010). While acknowledging the limitations of action by the state in of SI processes, we nevertheless believe it to be of paramount importance that public institutions promote and coordinate mixed networks of actors in order to achieve an effective and democratic accountability system. Indeed, the state should be brought back in because public institutions have the non-delegable role of overseeing the rights and duties of citizens and an equal distribution of opportunities, also protecting vulnerable people and minorities in their access to welfare. If innovation is not carried out by local institutions, the risk of weakening the social dimension of welfare in favour of efficiency is real (Saruis et al. 2019). Moreover, it should be borne in mind that at the local level there is also a right-wing, discriminatory and xenophobic civil society (see Castelli Gattinara, Froio, and Pirro 2021). Hence, local authorities steer innovation in order to promote and ensure protection of the rights of all. The role of local authorities is also to identify and avoid the risks of initiatives led by a majority civil society seeking ethnic coherence and exclusionary self-segregation. Finally, improvement in accountability can be achieved only through governance-led social innovation.

The growing body of literature on the co-production and co-creation of public services has been making it possible to reaffirm the role of the state in promoting citizen participation and community networking, i.e. its role in triggering social innovation processes, including improvements in social relations, structures of governance, greater collective empowerment (Moulaert et al. 2013; Pais, Polizzi, and Vitale 2019; Voorberg, Bekkers, and Tummers 2015). The term ‘co-production’, in fact:

describes, the potential relationships that could exist between the ‘regular’ producers (street-level police officers, school teachers, or health workers) and ‘clients’ who want to be transformed by the service into safer, better educated, or healthier persons. Coproduction is one way that synergy between what a government does and what citizens do can occur. (Ostrom 1996, 1079)

The concept of coproduction is rooted in the main models of public administration (Sorrentino, Sicilia, and Howlett 2018), each of them allocating specific roles to the public actors, citizens and civil society organizations, and civil servants (Sicilia et al. 2016). In this context, co-production has been identified as a driver of innovation in public services (Nesti 2018) – especially as regards the implementation of social and welfare policies (Campomori and Casula 2021). According to Lévesque (2013, 25–39), in fact, the various waves of public administration reforms have reflected different conceptions of how innovation should be pursued in the delivery of social and welfare services (see also the more recent classification of service delivery arrangements by Profeti and Tarditi 2021). Whilst in the traditional public administration (henceforth ‘TPA’) model – the paradigm dominant in public administration until the 1970s and the 1980s (Hartley 2005) and based on the separation between politics and administration (Weber 1968) – innovation in welfare and social policies refers to some large-scale, national and universal innovations, the New Public Management (henceforth ‘NPM’) model – characterized in the 2000s by the introduction of market-type principles and measures inspired by private-sector logics for the delivery of public services – pushed for the introduction of a radical innovation in organization forms and processes rather than content. Within the New Public Governance (henceforth ‘NPG’) model, innovation in the public administration takes mainly the form of inclusive and participatory governance, with an emphasis on collaborative partnerships between state and non-state actors (Sorrentino, De Marco, and Rossignoli 2016). However, the inclusion of a plurality of actors in policy-making does not diminish the importance of the state’s role as the enabler of innovation, as shown by recent studies (Campomori and Casula 2021; Casula, Leonardi, and Zancanaro 2022): the longer-lasting success of innovative organizations based on the involvement of citizens as co-producers of social services is made possible by institutional, legal and financial support from the state (Pais, Polizzi, and Vitale 2019; Vitale 2009, 172).

Following Ayob, Teasdale, and Fagan’s (2016, 649) argument, coproduction can therefore be conceptually linked to more radical models of SI, with the strong tradition of SI that can be seen as invoking ‘the synonymous narrative of restructured power relations through the engagement and empowerment of previously disadvantaged individuals and groups’ (Ibid.). Within the NPG model, according to Miquel, Cabeza, and Anglada (2013, 160), the analysis of SI in relation to governance can therefore be conceptualized by considering not only the internal dynamics of the actors promoting SI (‘governance as a framework for innovation’) but also the impact of social innovative strategies on the governance mechanisms and empowerment capacity of both collectivities and individuals (‘governance as a field for innovation’).

Given this theoretical background, we expect to find an innovative governance – in terms of both ‘governance as a framework for innovation’ and ‘governance as a field for innovation’ – which is more participatory in the presence of a social, institutional and political context favourable to innovation, and especially in the presence of proactive public actors.

2.3. A socially innovative oriented governance

The NPG model certainly assigns a role much more prominent than that envisaged by the TPA and NPM models to the non-profit sector and citizens in improving public policy and service delivery (Casula 2017): it emphasises ‘collaborative governance models based on reciprocal respect, interdependence and trust among the partners, rather than on contracting and competition as the animators of relations’ (Brock 2020, 260).

The NGP model therefore appears to be the one best suited to building a conceptual frame of the SI process. We are interested in investigating the specific arrangements that (should) take place when public and private actors are engaged in devising new ways to cope with new or old societal challenges and in assessing the specific features of a governance system informed by a reconfiguration of social relations. The body of scholarship that has proliferated since the late 1990s on the ‘thorny issue’ of the state/civil society relationship (Swyngedouw 2005) has partly taken a positive view of what have been called new governance systems, emphasizing the horizontal interactions among participants and the implicit improvement of democracy (Schmitter 2002), as well as the new opportunities for increasing effectiveness in policy making (Pestoff, Brandsen, and Verschuere 2012). This optimistic view is nevertheless challenged by a more sceptical one which highlights the Janus-faced nature of governance-beyond-the-state and warns that the imposition of market forces may finally impose the rules of the game (thus revealing a ‘democratic deficit’) (Swyngedouw 2005). Moreover, according to Aucoin (2012), the shift from NPM to NPG is likely to give rise to an era of new political governance where a ‘promiscuous partisanship’ may emerge and encourage a ‘dualistic view of politics in which those who are not allies of the government must be enemies’ (Aucoin 2012, 178): hence the NPG relationship and policy tools may be subjected to partisan manipulation at the expense of the broader public interest (Brock 2020, 262).

Taking into account this debate, which suggests not to assume an apologetic view of NPG and the new forms of governance beyond-the-state (Swyngedouw 2005), we now put forward our understanding of the main features of SI-oriented governance, while we also argue that a socially innovative governance system should be implemented in order to avoid possible pitfalls.

Firstly, socially innovative governance does not take the form of a mere contracting-out process whereby the public actor simply delegates the implementation of services to TSOs, mainly for having granted a budget saving in time of fiscal stress: socially innovative governance involves much more than the production and delivery of public services (Cheng 2018). Moreover, civil society should never replace the local state in remedying state failures; rather, it should be supplementary (Gerometta, Haussermann, and Longo 2005).

In our view, the key for innovation in governance is improvement of a shared commitment primarily when setting the problem. In fact, SI cannot ignore the representation of new ways to frame and understand the issues at stake. Both state and non-state actors can provide significant knowledge and points of views to reach a non-simplistic definition of problems and subsequent actions: while civil society can count on close proximity to potential users and on a remarkable flexibility of action, the public actor should maintain its role in regard to citizens, i.e. ‘guaranteeing quality of services equity, fairness and access. In synthesis full and equal citizenship’ (Oosterlynkck et al. 2013).

More specifically, socially innovative governance means fostering a bottom-up approach that springs from open discussion among state and non-state actors, experts, stakeholders, potential target groups and beneficiaries of the policy. Discussion should disentangle the diverse aspects of the issue at stake, as well as bring out the various interests and policy constraints (e.g. budget constraints, consensus constraints, or technical constraints) embedded in the issue (this is the framing activity which includes (co)-setting the problem). The literature on co-design and co-governance has already pointed to a new role for non-profit organizations as structures mediating between the political order and individual lives (Cheng 2018, 2); and the interest of scholars has progressively shifted from the TSOs providing public services to TSOs planning and designing public services. Here we take a step forward by highlighting the role of non-profit organizations and citizens already in co-framing the issues at stake.

3. A three-step framework for designing socially innovative practices

In this section we describe a methodological framework in which both to analyse social innovation processes and to design a socially innovative governance system. The framework is based on three consequential steps: (i) a preparatory phase which comprises an in-depth contextual analysis conducted from a synchronic and diachronic perspective; (ii) an analytical phase where a sample of best practices related to the policy at stake are scrutinized in order to grasp ‘what works’ (Coletti 2013, 94); (iii) a co-planning phase which consists of a co-framing and co-creation activity. This third step will initially involve key actors identified in the first step and then other significant actors whose names and profiles emerge from a snowball sampling process. In addition to being characterized by use of the various qualitative techniques that will be detailed in the next section, each of these three steps has a specific theoretical foundation.

Regarding the preparatory phase, the Institutional Collective Action Framework (henceforth ‘ICA’) is a good theoretical basis for a preliminary contextual analysis of how and where social-innovation practices arise. The ICA Framework, in fact, interprets the emergence of cooperative implementation arrangements as resulting from dependence on specific contextual and institutional factors that can reduce the contractual risks and transaction costs of cooperation for local actors (Feiock 2007, 58). More in detail, the latter can be reduced by specific community characteristics and formal or informal institutional arrangements: for instance, homogeneity in pre-existent networks of relationships among local agents (Feiock 2007, 2009), and/or the presence of regional diffusion mechanisms and regional governance bodies (Casula 2020). These theoretical foundations of the ICA Framework find fertile ground for application in the field of SI. As explained in previous sections, in fact, the emergence of social innovation practices in Europe is context-dependent (Moulaert et al. 2005; Benneworth and Cunha 2015). This context-dependence of social innovation practices has for example been found on the ‘social innovation framework model’ identified by the TEPSIE project that we described in section 2.1 (Krlev, Bund, and Mildenberger 2014; Bund et al. 2015). In particular, what we call the ‘preparatory phase’ implies recognition of what TEPSIE terms the framework conditions, i.e. the innovation capacity or the innovation potential of a specific context (city or region). These conditions have then be further detailed in four more specific categories (Krlev, Bund, and Mildenberger 2014) which allow a more effective operationalization. They are: (1) the institutional framework, i.e. the set of values, norms and laws that govern human actions at a social level, including the political culture of a context; (2) the societal climate framework ‘which covers attitudes towards change and openness to the development of (social) innovation just as civic engagement in political and social life or the existence of a shared set of needs and awareness within society for the latter’ (Krlev, Bund, and Mildenberger 2014, 204); (3) the resources framework, including tangible resources like money or technology and less tangible ones like social capital and explicit and tacit knowledge (Polany 1966); and (4) the political framework, which involves for instance an analysis of the incentives and actions by the political system aimed at encouraging social innovation.

The second step starts from the assumption that policy-makers usually explore the adoption of innovations that have been previously developed in other contexts with the intention to transfer programmes and practices experienced elsewhere into their local context policies. The phenomenon of ‘policy diffusion’ has been described by the literature on policy learning which, by focusing on classic studies relating to policy transfer, analyses how a policy is transferred from one country to another, or from one city to another within the same country: that is, what Rose calls ‘lesson drawing’ (Rose 1991, 1993, 2004). Notwithstanding the consolidation of this literature – which now includes both the prescriptive and the descriptive dimension of what it is possible to learn from other experiences – the policy transfer approach has some limitations, the most important of them being that it does not take due account of the fact that the transfer of an innovation is neither an easy nor an automatic process. Accordingly, Bardach (2003, 2008) prefers to use the concept of ‘smart practices’ (rather than ‘best practices’), since while the adjective ‘best’ implies that a practice is better than other ones, the goal should be to make the most of hidden opportunities. Moreover, in order to increase the ability of policy makers, practitioners and policy analysts to ‘learn how to learn’, Coletti (2013) has recently developed a set of guidelines for policy designers to enhance their capacity to transfer an innovation from a source case – that is, where an innovation has been invented – to a target case, that is, where the innovation has to be transferred. Coletti recognizes that during the transfer process each ‘smart policy designer has to pinch ideas and creatively adapt them to the chosen target context’ (p.93). In particular, after screening the practices that seem to work better and comparing among two or more successful cases, Coletti suggests collecting evidence of ‘what works’ in these source cases and then analyse the policy process. These shared preliminary reflections should ‘safeguard the policy designer against the risks involved with a mindless implementation and replication of a policy’ (93). Later, they help policy designers to understand if and how it is possible to transfer the innovation, in particular as regards the design of process features according to the target context.

Co-framing and co-creation are the final steps of our framework, and their purpose is to create genuine and meaningful interaction among key stakeholders. A major advantage of the use of co-creation is that the cooperation among all the people involved (the quadruple helix approach to SI represented by government, citizens, industry and academia) guarantees a plurality of information and perspectives for improving policy design (Peters 2015; Wollman 2007). Co-creation is therefore much more than just participation of the target group because it includes a commitment by the group throughout the policy design process, not just a few sporadic stakeholder consultation meetings. In order to be effective, co-creation must not be improvised and therefore requires some strategies to be followed, which Arnim Wieck (2016)² has identified as the following: (i) an up-front clarification of objectives and processes, since co-creation processes need to be carefully designed, with clear objectives (expected outcomes) and processes (who collaborates with whom, when, and how); (ii) the identification of relevant stakeholders and the use of a well-balanced engagement throughout; (iii) the use of neutral professional facilitators that should enable a just and open engagement process; (iv) the choice of an appropriate process;³ (v) the presence of sufficient and reasonable resources for stakeholder engagement processes, facilitators, and experts in co-creation; (vi) the use of a formative evaluation to assess whether the co-creation process had made a difference in the complex problem. Applied to the field of SI, this approach guarantees a high degree of reflexivity on the part of all the parties engaged in the co-creation of a socially innovative practice. In fact, in order to guarantee social inclusion, marginalized groups participate together with other stakeholders, and they take responsibility in SI ‘for the construction of arenas in which alternative definitions of social problems can be given systematic form and made accessible to a broader public’ (Vicari Haddock 2013, 427). Both stakeholders and marginalized groups bring different needs, interests, capacities, and resources to the table, and the co-creation process should take this diversity into account. At the same time, it is important to bear in mind that the effectiveness of the co-creation process also depends on the framework conditions mentioned above (Casula 2015), and this tells us that the process will not have the same effectiveness or feasibility everywhere. In this regard, Katrin Prager (2016)⁴ argues that co-creation is more likely to be successful if there is a shared problem that has some degree of urgency, if stakeholders have an interest in solving it, and if they feel that they cannot solve it on their own.

The next section shows how this methodological framework has been applied to the case of the integration of refugees and asylum seekers at the local level. In light of this illustrative example, it should become clearer how the three consequential steps integrate with each other in designing innovative social practices. At the same time, the example also reveals the possible obstacles to the construction of an innovative social governance process.

4. An application of the method to refugee integration policies

The above-proposed three-step method has been applied to the area of refugee integration policies in five cities, i.e. Bologna and Parma (Italy), Berlin (Germany), Vienna (Austria) and Ljubljana (Slovenia). The concrete experimentation of this framework is part of the implementation of the European project ‘Integrating Refugees in Society and Labour Market through Social Innovation’ (SIforREF), within which the framework was also developed. Refugee integration is a paradigmatic case for testing a socially innovative governance system. It is in fact a typical intractable and wicked problem, which resists ‘traditional’ solutions and approaches, since solutions are often unclear and any intervention may have unforeseen consequences (Peters 2015, 30). In addition, migration is a highly politicized issue, likely to become divisive and a source of bitter political and social conflicts (see Bazurli, Campomori, and Casula 2020). Usually, while NGOs and social movements strive to improve rights and integration opportunities for refugees, policy-makers are much more reticent if not decidedly averse to refugees. The main objective of the SIforREF project is to overcome the current short-termism of reception policies and address the issue of the medium- or long-term integration of refugees. The starting point of SIforREF draws on the theorization of Schon and Rein (1994) according to which problem-setting is the crucial phase of the policy-making process, even more so when an issue is highly controversial. Socially innovative governance cannot be separated from an attempt to create the conditions for a definition of the problem in which the aspects linked to stakes, ideologies, and the search for consensus are at least explicit and where it is possible to play uncovered cards, so to speak. The aim of socially innovative governance is to enable a new and shared way to set the problem, i.e. to yield a new understanding of the problem itself. The composition of the consortium reflects the objective of facilitating an innovative governance process: SIforREF’ s partnership, in fact, comprises four non-profit organizations working for and with migrants, four public actors, and three research institutions experienced in migration. From the outset, we wanted to enable a dialogue and a genuine debate among actors with different rationales and logics of action. These partners were the forerunners for engaging other actors and experimenting with our methodology.

The purpose of the first preparatory phase was to gain familiarity with the contexts in which cooperative governance arrangements arise and to obtain in-depth understanding of the local ‘battleground’ of migration (Campomori and Ambrosini 2020). Remembering that in the sociology of institutions and state agencies there is a principle of symmetry that requires analysts in the field of migration to study not only what is done but also what is not done (or what is avoided), recent studies have in fact shown that European states have developed different policies for assimilation and integration by developing different models based on specific institutional and contextual factors (see, King, Le Galès, and Vitale 2017). We believe that this first phase is the foundation of the SI process, since socially innovative practices are deeply rooted in specific contexts with specific sets of conditions and different contextual and institutional factors that need to be comprehensively understood. In each city included in the project existing policy networks (including positioning of state and non-state actors) has been investigated, as well as power relations (politics), social and political climate, and the overall policymaking process on refugees in a diachronic perspective. During the three months of this first step of the methodology, the project’s researchers interviewed at least ten key informants per city and analysed policy sheets, local online newspapers and databases with statistics on refugees, the labour market, polling data, the unemployment rate, and traditional vitality of voluntary associations. The construction of the questionnaire, as well as the choice of the people most interesting to interview in each context, was the result of intense discussion among the project partners, who made use of their background and knowledge. Researchers conducted the interviews. The presence in the partnership of both public actors and civil society enormously simplified the task of gaining trust in the researchers, thus decreasing the transaction cost due to diffidence, scepticism and, sometimes, time constraints. It is in fact well recognized that the support of civil-society actors has facilitated the integration and well-being of migrants (Garkisch, Heidingsfelder, and Beckmann 2017). Moreover, this phase was crucial for weaving together a network of contacts that we engaged in the following steps of the project.

The output from this step was a detailed analysis of the refugee reception and integration policy framework: in each context, we devoted especial attention to clarifying the role and the positioning of policy-makers, practitioners and non-state actors; the local governance dynamics have also been analysed.

Some evidence on existing social innovative practices already emerged in this phase. It was collected and collated for use in the following step. In fact, the second phase mapped a number of recognized (by key informants) best practices and sought to grasp ‘what worked’ in the source case (Coletti 2013, 93) and which political, social or territorial conditions made the emergence of these practices possible. In order to achieve this result, project partners from civil-society and from public authorities worked together on a template in which all the most relevant information on the best practices was inserted and then thoroughly discussed. The goal of this step was not to transfer a practice from one context to another, but instead to understand, and to make knowable outside the partnership, the key features of best practices, as well as why they are more likely to work in a specific context.

The design phase (step three of the methodology) envisaged activation of the main actors involved in refugee integration in each of the cities surveyed. The group comprised at least ten people and consisted of policy-makers, practitioners, stakeholders, refugees, and social workers. These actors were invited to engage in recurrent rounds of discussion supported by professional facilitators, in a setting which resembled that of deliberative democracy experiments (Gastil and Levine 2005). In the first round, participants were asked to share their points of view, ideas, and perceptions on refugee reception in their city, according to their institutional or social role. In the second round, participants were invited to suggest possible solutions for the difficulties experienced, such as language barriers, lack of skills of refugees, excessive bureaucracy, and lack of childcare for women. In the third round, one or more best practices implemented in each city were presented and peer reviewed by the participants, the purpose being to prompt further discussion on the following points: (i) what is a best practice in the realm of refugee integration; (ii) what are the ingredients of success; (iii) what are the possible obstacles to implementation; and (iv) what are the outcomes yielded by these practices. During the workshops, attention was constantly paid to how the issue could be framed, taking account of the different perspectives and backgrounds of the participants (e.g. policy-makers have to cope with consensus-building, employers need people with language skills, TSOs and voluntary associations want to augment refugees’ rights and opportunities). To this end, role-games were also used. The active involvement of civil-society actors and refugees was particularly important for analysing the specific problems encountered by practitioners in their everyday activities. civil-society representatives, for instance, gave useful advice on how to overcome the language barriers that limit participation by refugees, how to effectively involve civil society in mentoring integration processes (for instance, the home-sharing schemes widespread in Europe), how to provide better coaching in job centres, how to approach proactive refugees vs. refugees who need more support because of their limited individual resources.

4.1. Products of co-creation process and discussion

This co-framing and co-designing activity produced two main sequential outputs: the building of a decision support tool with which to design and implement inclusive and innovative practices in the realm of refugees; the piloting, for about six months, of this tool through the implementation of 7 pilot actions in the five cities involved. We focus here on the pilot activities that constitute one of the most original products of the project in relation to the SI process. In their development we can also observe the strengths and weaknesses of the implementation of a socially innovative governance system. Pilots were designed according to the local challenges identified in the previous steps of the project, and their aim was to improve the labour-market and/or social integration of refugees and asylum seekers. Some of the pilots built on existing activities, to which they added some specific services based on the needs that emerged during the co-design activity. These included pilots in Vienna and Berlin (see Table 1 for a summary of all the pilot actions). In particular, as regards Vienna, the pilot called Magdas Hotel set up a specific training programme, for refugees working in the hotel, on conflict management and the development of intercultural competences. The pilot called Volunteers for Volunteers was based on the Community Buddies service set up by Caritas in 2012 and aimed to create a training curriculum for new volunteers by using experienced volunteers. In Berlin, the two pilots focused on labour-market integration and particularly concerned refugee women aged over 35 with low German language skills and interested in exploring self-employment. The Italian and Slovenian pilots, instead, did not arise from follow-ups on other projects. In Bologna and Parma the aim was to improve the social inclusion of refugees and to reduce prejudice against them in the host society. The Bologna pilot (Welfare community management), carried out by ASP (Public Company for Personal Services), worked on the creation of events in which refugees could feel like protagonists and deploy their resources and skills; the pilot of Parma (Refugees as Social Caretakers) was managed by the Municipality of Parma together with a local NGO, and it sought to foster positive neighbourly dynamics between refugees and natives. In Ljubljana, the pilot (Connecting refugees with the labour market, education and craft) worked with 25 refugees, investigating their needs, resources and desires. From the results of this investigation the pilot, carried out by an NGO, implemented a communication platform advising refugees on various possibilities of inclusion in the labour market, connecting and providing info on additional educational, vocational options.

Table 1. Summary of pilot actions.

Cities, countries	Parma, Italy	Ljubljana, Slovenia	Bologna, Italy	Vienna, Austria	Vienna, Austria	Berlin, Germany	Berlin, Germany
Name of the pilot	Refugees as Social Caretakers	Connecting Refugees with the Labour Market, Education and Craft	Welfare Community Management	Magdas Hotel: soft skills for apprentices	Volunteers for Volunteers	Incubator Cafe for Refugee Women	Refugee Women and Pathways into the Labour Market
Summary of the pilot	Four refugees, trained to become ‘social caretakers’ (‘portinai sociali’), lived in small apartments in multicultural residential areas and offered their coinhabitants both practical help (taking care of sorted waste collection; helping people in need with their everyday tasks; small condominium maintenance jobs) and social support (conflict prevention and mediation; approaching the other foreign residents and fostering their integration; organizing social events and small-scale services).	The main idea of the pilot was to work with 25 refugees to investigate their needs, expectations and skills; after the interviews, joint meetings were organized to share ideas on a common communication platform. Its purpose was to connect refugees with the local labour market (LM) through advisers (also other refugees). Hence the main purpose was to form a communication platform that would assist refugees with knowledge about various possibilities of inclusion in the LM, connecting and providing information about additional educational, vocational and arti options.	The aim of the pilot was to contribute – through community activity and social inclusion workshops – to the growth of a culture of relationship and participation among citizens, asylum seekers and refugees living in the city of Bologna. The pilot developed community workshops in 4 areas of the city.	A new educational line for apprentices at Magdas Hotel was created. This educational line was tailored to apprentices with migrant experience. A new curriculum was formed and through two training sessions was trialled and revised. The key feature of this pilot was that also in the creation and implementation of the curriculum people with migrant experience were part of the team. The aim was to give apprentices training in soft skills – with a focus on conflict resolution and intercultural communication – so that on the one hand, they had better chances in the labour market after finishing their apprenticeships; on the other, more options for action in difficult circumstances, hence an active part in social cohesion and integration.	Twelve volunteers (mostly with migrant or refugee backgrounds) were trained. They received information on ‘labour market integration of refugees and migrants’ during five training workshops. Six experienced volunteers from ‘Grätzeleltern / Community Buddies’, designed, organized and implemented the training programme. After completing the course, the trained volunteers passed on the knowledge they had acquired to migrants and refugees in Vienna, and directly improved their situations.	‘Self-employment & me’ orientated opportunities for self-employment in Berlin to women with refugee status who were not being reached by mainstream immigrant services. This orientation programme brought 2 groups of 12 refugee women together to explore empowerment, resilience, business ideation, and technical requirements of self-employment. The 10 – week programme comprised one weekly 3.5hr workshop, and one hour per week of one-to-one mentorship. The first cohort received training and mentorship exclusively in Arabic. The second cohort received the training in in Farsi.	Project devoted to job-seekers (men and women) supported and accompanied through all the phases of the job search (offering them individual consultations), finding suitable job offers, creating application portfolios, accompanying them in the application process, preparing them for the interview, and also supporting them after their integration in the labour market. Through this pilot, closer attention was paid to the obstacles and hurdles that refugee women face when integrating into the German labour market.

Both the development of the decision support tool and the conduct of the pilots showed some differences related to specific local contexts. In particular, where the framework conditions were not particularly supportive, for example because the local government did not have an open attitude towards refugees, it was difficult to involve policy-makers, and this limited the innovativeness of the governance process. Consequently, the pilots in these cases were developed in complete independence from public actors, with the risk of encountering sustainability problems in the long term. This situation occurred in particular in Ljubljana. More generally, however, in all the cities – even where there was an institutional partner as a member of the project consortium – the involvement of political actors was tricky, albeit strongly requested. Politicians, in fact, preferred to delegate officials or managers to take part to co-creation round tables, as if to signal that this type of activity was not considered important or that it took time away from other matters. Furthermore, the process that took place in the project revealed how important it was for a local context to be able to count on TSOs that were already strong, structured and oriented towards SI, as in the case of Caritas Vienna, which coordinated the two pilots of the city. Also in this case the sustainability of the actions implemented was more certain when a strong TSO was present and active in the process and, at the same time, its legitimacy made it easier to mobilize political actors.

5. Conclusion and discussion

This article draws on the growing literature related to the changing operational paradigms in public administration reforms. Throughout this process, the relation between public and non-profit sectors undergoes many changes when new public policies that governments decide to develop to deal with emerging societal challenges are implemented at the sub-national level, the most visible being contracting-out in order to achieve policy goals (under NPM) and a multi-actor policy making, including co-design and co-production (under NPG) (Brock 2020). The specific theoretical effort of the article has been to connect the literature on the new governance arrangements, of which we have highlighted strengths and possible weaknesses, with the concept of SI, in an attempt to build a more robust analytical framework for research on SI itself. We explicitly linked SI to a governance innovation, understood as a changing relation between state and non-state actors, including citizens, target groups and beneficiaries of public policies.

SI studies would benefit from our methodology in two main respects. Firstly, it can be used to assess if and to what extent non-state actors are actually involved in horizontal and cooperative relations with public actors for the delivery of public services; in other words, it can be used to check if private and public actors make an effort to jointly frame the problem and if there are real opportunities for non-profit organizations and citizens to influence policy development. Secondly, the proposed methodology can guide scholars in investigating which non-state actors are best suited to socially innovative governance within a public policy process (especially during the implementation stage). In this regard, a possible research question might be this: what are the features of civil society organizations best suited to enabling socially innovative governance? For instance, how important is it for non-profit organizations to have a high degree of reflexivity, a staff skilled and trained in problem framing, and solid experience in entering into relations with public actors? As also argued by Pestoff (2012, 4), not all non-profit service providers are equally prepared to play a leading role in the development of public financed services: while some organizations can significantly contribute to enhancing and promoting democratic governance, others will probably play a less prominent role.

Methodological considerations also arise when studying SI informed by our approach. Two issues seem of particular relevance.

Firstly, while in the policy-making process the public actor is present by default, third-sector participation is usually subject to a request by the public actor itself. Hence the question arises as to which organizations are more likely to be actively involved in the policy-making process and thus in the co-creation activity. In other words, could it be the case that a number of TSOs making a valuable contribution to the issue at stake are actually not involved at all because they do not have the ‘right’ features, according to public actors, or they are not sufficiently visible in the public sphere?

Secondly, and consequently on the first point, governance inevitably displays some degree of politicization (more pronounced on some issues than on others), since it is part of a policy–making process where conflicting interests confront and clash. In other words, it may happen that political governance strongly influences the relation between public and non-profit sector from the very beginning of the process. This issue has been raised also by Brock (2020, 265) as a finding of her research on policy hubs and innovation laboratories (henceforth ‘PILs’) in Canada. She points out that non-profit sector partners in PILs may become part of the process of politicization of the public sector. In this regard, the case study on refugee integration offers further material for reflection. As we specified in the previous section, the SIforREF’s partnership consists of public actors and non-profit organizations committed to refugee reception and integration. Within the project, this partnership found a ‘container’ in which to experiment with bargaining informed by trust and mutual learning, moving away from the classic result-oriented political bargaining. We maintain that the main reasons for this result were the opportunity to undergo a process where reflexivity had a recognizable space and time, where time for dialogue was not curtailed, and where partisan interests could be temporarily set aside. Indeed, the setting provided by the project made possible a relational climate more similar to a deliberative process than to a political negotiation process.

To conclude, this article has proposed a more robust analytical framework in which to evaluate the growing number of social innovative practices that are implemented in different contexts and policy sectors in order to deal better with new grand challenges. Through the presentation and the application of a specific methodology, this article has provided elements useful (i) for guiding public administration scholars and practitioners in investigating which features must be improved in order to enhance socially innovative governance within a public policy process; and (ii) for assessing if and to what extent civil society actors are actually involved in horizontal and cooperative relations with public actors when new multi-actor implementation arrangements to deal better with emerging societal challenges are established. Our data supported the general hypothesis that social innovative governance – in terms of both ‘governance as a framework for innovation’ and ‘governance as a field for innovation’ – requires a social, institutional and political context favourable to innovation, which is difficult to achieve in the absence of a proactive role of public actors.

Acknowledgment

The authors contributed equally to this work. The authors would like to thank Innovation: The European Journal of Social Science Research’s editors and the anonymous reviewers for their valuable advice and comments on previous versions of this article.

Disclosure statement

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

Additional information

Funding

The article was supported by the research made within the project SIforREF, Integrating Refugees in Society and Labour Market through Social Innovation (project index number CE1527) https://www.interreg-central.eu/Content.Node/SiforREF.html. The project is funded by the Programme Interreg Central Europe and it addresses the risk of marginalization of refugees after the reception phases in 5 cities (Berlin, Vienna, Bologna, Parma, Ljubljana). SIforREF used the approach of Social Innovation for implementing 7 pilot actions on refugee integration. Findings and results of the project are uploaded on project website.

Notes on contributors

Francesca Campomori

Francesca Campomori is Associate Professor in Political Science at the Department of Philosophy and Cultural Heritage (University of Venice Cà Foscari), where she teaches Social Policies and Public Policy. Her main research interests are in the field of immigration policies, especially in a local level perspective, social policies and social innovation. Currently, she is leading a three years project called SIforREF Integrating Refugees in Society and Labour Market through Social Innovation CE1527 https://www.interreg-central.eu/Content.Node/SiforREF.html. The project is funded by the programme INTERREG Central Europe.

Mattia Casula

Mattia Casula is currently a Research Fellow in Political Science at the University of Bologna, where he teaches Public Administration and Organization Theory. His research interests and publications are in the field of public policy and administration, with a focus on the study of implementation dynamics in different multi-level and multi-actor organizational settings. Former member of the Board of the Regional Studies Association, he is currently member of the Executive Committee of the Italian Political Science Association and co-coordinator of its Standing Group “Regional Studies and Local Policies”.

Notes

1 The other two dimensions are the content/product dimension, i.e. the satisfaction of unsatisfied human needs; and the empowerment or socio-political transformation dimension, related to increasing the socio-political capability and access to resources needed to enhance rights to the satisfaction of human needs and participation.

2 Cfr. Arnim Wieck (2016): Eight strategies for co-creation. Available at: https://i2insights.org/2016/05/12/eight-strategies-for-co-creation/.

3 According to Wieck (ibid.), there is a wide range of co-creation processes, including: Listening sessions that allow stakeholders to air their concerns, perspectives and ideas; Discussion sessions among stakeholder groups (which can be diverse or homogeneous) aimed at exchange and mutual understanding; Collaborative sessions on project deliverables; Elicitation sessions to receive feedback on deliverables. Interactions can be via interviews, surveys, focus groups, walking audit workshops, or other means; engagement can be virtual or face-to-face.

4 Cf. Katrin Prager (2016): Is co-creation more than participation? URL: https://i2insights.org/2016/07/28/co-creation-or-participation/ [05.01.2022].