Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

The aim of this book has been threefold: first, to describe the evolution of research-policy dialogues on migrant integration in Europe at different levels – the local, the national and the EU level; secondly, to find patterns of convergence and divergence in this development; and third, if we see such patterns, to explain why these appear. In our search for patterns and explanations, we come back to the conceptualisation discussed in the introductory chapter. There we formulated three basic empirical questions: (1) What forms of research-policy dialogues do we find empirically? (2) What type of knowledge is brought into these dialogues? (3) What use is made of that knowledge?

We are also interested in the interrelationship between our three major concepts: research-policy dialogue structures, knowledge utilisation and knowledge production. These influence one another, and this generates questions such as: how do specific dialogue structures affect knowledge utilisation, and vice versa? Do certain forms of knowledge utilisation lead to specific dialogue structures? What are the effects of dialogue structures on knowledge production, and do certain forms of knowledge production lead to specific dialogue structures? And to what extent are these modes of utilisation and production of knowledge a result of, and embedded in, national traditions of research-policy relations – or, more broadly, in dialogues between science and society?

One specific question was added to each of the foregoing ones: has the politicisation of migrant integration, which seems to have occurred in many parts of Europe over the last decade or so, had an impact on dialogue structures, knowledge utilisation and knowledge production and their interrelationships? And, if so, what has been the nature of that impact?

2 Forms of Dialogue Structures, of Knowledge Use and Knowledge Production

2.1 Forms of Research-Policy Dialogue Structures

The 15 preceding chapters brought together empirical studies of research-policy dialogues from a variety of contexts. In Part I of this book several specific forms of dialogue and their functions have been analysed, such as the use of statistics by policymakers, the role of knowledge production inside government institutions, the impact of national integration models on research and policymaking, and the role of expert committees in shaping integration policy. Most of these phenomena have been studied on a comparative basis, which has revealed similarities, but also differences between European countries in their ways of handling the interrelationships between knowledge production and knowledge utilisation, between research and policymaking. As the role of the European Union in this field of policymaking has become more prominent recently, three chapters in Part I were dedicated to research-policy dialogues at the European level, including one dealing specifically with local policymaking and the role the EU has been playing here.

Part II of the book includes analyses of research-policy dialogues in a number of ‘old’ immigration countries (Austria, Denmark, Germany, the Netherlands and the United Kingdom), as well as Italy as a relatively ‘new’ immigration country, and Poland as a country that has only just begun to experience immigration, after having witnessed much more sizeable emigration and transmigration. This part of the book also includes an analysis of the increasingly relevant research-policy dialogues at the EU level.

All chapters indicate that the emergence, development and effects of research-policy dialogues are very diverse. Long-standing forms of dialogue structures, both at the national and the local level, have often changed significantly over the last decade, but in different ways and at different paces. New dialogues have emerged, particularly, but not exclusively at the EU level. All of this has added undeniably to the diversity of the interrelationships between knowledge production and knowledge utilisation, of research-policy structures, and, even more broadly, of science-society dialogues. It has also become clear that integration policies increasingly are a multi-level issue of governance. Not only the role of the EU has become more prominent, but also that of local government. This again has contributed to more diversity in research-policy dialogues and to the development of new coalitions, for example between the European Commission, cities and NGOs.

2.2 Knowledge Utilisation

The empirical cases brought together in this book all draw on Boswell’s distinction between instrumental and symbolic modes of knowledge utilisation (Boswell 2009; see also Chap. 2 in this book). Instrumental knowledge utilisation refers to the use of knowledge directly in order to rationally adjust policy outputs, as is assumed in the term ‘evidence-based policymaking’. This function of knowledge is found regularly in different phases of the policy cycle: either at the stage of conceptualising and framing new policies, or, more often, in developing concrete instruments for existing policies and in monitoring and evaluating policy action. In recent years, this instrumental and evaluative use of research has become clearly visible at the EU level. This may be to compensate for the fact that the European Commission has only limited competences in this policy area, so that its credibility depends on its ability to mobilise relevant knowledge and instruments. The same phenomenon may be observed at the local level, where a systematic exchange of experience and good practice in using policy instruments has become a major element of knowledge utilisation.

Nevertheless, the other mode distinguished by Boswell, symbolic knowledge utilisation, appears to have gained in importance. This refers to more indirect ways of using knowledge, either in substantiating already-decided policy choices or in legitimising policy actors. The chapters on countries that do have a longer policy tradition in the field of migrant integration confirm such symbolic use of knowledge. This becomes evident, for example, in the selective use of research-policy dialogue structures and their products by policy actors, as a good number of government-initiated expert commissions have experienced (see also Boswell and Hunter 2014). Such specific use of knowledge can be facilitated by specific structures that produce knowledge, for example within ministries that set up their own in-house research departments. Local authorities may do the same.

Some of the contributions to this book suggest that we should go beyond Boswell’s conceptualisation, for two reasons. First, Boswell’s distinction between instrumental and symbolic use of knowledge suggests that these two exclude each other. In reality, however, the same input of knowledge may have an instrumental and a symbolic function at the same time, or such input may have a symbolic function initially, and later become instrumental as well. This has been observed in the analysis of several consultative expert commissions in the UK, Germany and the Netherlands.

Secondly, Boswell’s dichotomy seems to exclude a situation where no use at all is being made of knowledge, even if it is produced at the request of policymakers and funded by them. In her chapter on Austria Borkert observes the phenomenon of ‘knowledge shelving’, whereby knowledge is stored for possible later use. This phenomenon can also be found at the European level. The European Commission has invested significantly in research on migration and integration (through the Framework Programmes of DG Research, but also through the European Refugee Fund, the European Integration Fund and the European Social Fund), but with little direct use. Much of what is produced does not filter through into EU policy, partly due to the European Commission’s lack of competencies in this field.

2.3 Knowledge Production

The third concrete empirical question in this book relates to changes in knowledge production and developments within the field of migration research itself. From all historical observations in this book it is clear that, initially, research was strongly embedded in national contexts. Consequently, the central issues and concepts are framed rather differently from one country to another, as a comparison of different chapters illustrates. In the UK, for example, the ‘Race Relations’ and, later, the ‘Community Cohesion’ frames have dominated the research field. In Germany this was the case for the ‘Guest Worker’ frame, in the Netherlands for the ‘Ethnic Minorities’ frame, while the ‘Republican’ frame was dominant in French research. In exceptional cases, when policies were first being developed, researchers, policymakers and the general public shared such a framing, at least for a certain period. Under such conditions a ‘co-production of models’ could develop: the same model was then used as an analytical tool by scientists and as a conceptual tool in policymaking (see also Entzinger and Scholten 2014). In most cases, however, the framing by researchers within one country was more diverse, also reflecting different scientific and methodological traditions. If there was a dominant concept at all in research it did not necessarily correspond with the way policymakers had framed the issue. As a general rule we may say that the closer the contacts between researchers and policymakers the more likely they are to use shared frames of analysis. Under such circumstances research activities often tend to be directly funded by the government or other stakeholders, and less so by universities or research councils.

Comparing our case studies, a general picture of the development of research-policy dialogues over time emerges. Roughly until the turn of the millennium, migration and integration research was strongly embedded within national contexts, not only in terms of the framing of questions, but also in the selection of questions to be researched, and in terms of funding. In the 1990s, research institutes dealing specifically with migration and integration began to emerge in many West European countries, mainly (but not exclusively) within universities. Initially, most of these institutes were primarily oriented towards the national context. How relations between research and policies developed in such national contexts differed significantly. Usually, however, politics and policy were the partners that defined and developed the relationship, if there was any, and not academia.

Later this picture began to change. The most important development has been the entrance of the European Union, and specifically the European Commission, as a policymaker and a funder of research. The European Commission has commissioned an increasing number of studies since international migration was declared a topic of communitarian policymaking in the Amsterdam Treaty of 1997. The field of integration followed in 2003 after the Commission’s Communication on Immigration, Integration and Employment had been politically accepted (see Chaps. 3 and 7 for a more complete overview of EU involvement in this research area).

Partly anticipating this development at the EU-level and partly in its wake, the research world started to organise itself across national borders (but clearly within the EU, or rather the European Economic Area). The IMISCOE Network of Excellence (2004–2010), funded by the European Commission, and its independent successor the IMISCOE Research Network, are the most prominent examples of this. IMISCOE’s programme is predicated on systematic cross-national comparative research, beyond ‘methodological nationalism’ (see Penninx et al. 2006). Some observers, however, wonder if the heavy reliance on EU-funding does not introduce a new dependency of researchers on the definition of migrant integration as a European problem (see Geddes and Scholten 2014).

3 The Effects of Politicisation

In the introductory chapter we formulated a number of hypotheses about the effects of politicisation on (the development of) dialogue structures, knowledge utilization and knowledge production. These assumptions speak to a broader literature on the role of science in contemporary late-modern society. Various scholars have argued that society is moving beyond the traditional model of ‘science speaking truth to power’, based on a powerful belief in the rational use of scientific knowledge (Wildavsky 1979). After the ‘scientification of politics’ that characterised this rational approach to policymaking, the late-modern or risk society is experiencing a simultaneous ‘politicisation of science’ in which science’s credibility is put on the line and scientific disagreements increasingly surface (Beck 1992; Gieryn 1999; Weingart 1999).

We have taken migrant integration as a case study for testing such presumed transformations. Migrant integration is increasingly perceived as a social risk and has become strongly politicised in many European countries. Throughout Europe, immigration and migrant incorporation have risen on national political agendas. Multicultural policies – whatever these may entail – have been discarded as a failure in many places, even in places where they have never existed at all. This (perceived) failure of rational efforts to establish a multicultural society and the political saliency of migrant integration have made this issue one of the most challenging social risks in the eyes of many Europeans (Scholten and Verbeek 2014).

Our first hypothesis, based on a review of the literature on migrant integration research and policymaking, was that the politicisation of migrant integration would contribute to a de-institutionalisation of existing research-policy dialogues. This speaks to a broader debate in the migration literature on the strongly policy-oriented nature of research, especially in the early years of the development of this field. Favell (2003), for example, pointed to a policy-oriented habitus of research and the consequences thereof. He argued that such an orientation would favour certain types of research and ignore others. Scholten (2011) showed that, especially in the Netherlands, a strongly institutionalised relationship between researchers and policymakers persisted for a long period: researchers were often directly involved in policymaking, but as a consequence they were also dependent on the policy context. In the science studies literature, this type of relationship has been described as the technocratic mode of research-policy dialogues (Weiss 1977; Hoppe 2005), with direct relations between researchers and policymakers and often also a central role for these researchers in policymaking. Elaborating this line of thinking, more recent work has argued that the politicisation of the debate on migrant integration in many countries has challenged this technocratic model and re-established a firm political primacy (see, for instance, Scholten and Verbeek 2014). In various countries this renewed political primacy has been associated with controversy over the involvement of researchers in this area, as in France where ethnic statistics are contested, or in the Netherlands where a public debate has taken place on the credibility of migration scholars. So, our first hypothesis assumes that politicisation leads to a de-institutionalisation of research-policy relations, which will become less direct, more open to diverse participants and more ad-hoc. By contrast, institutional relations will persist in places with a relatively low level of politicisation.

Our second hypothesis focuses on knowledge utilisation. Whilst instrumental knowledge utilisation involves the direct use of knowledge in policy formulation and political decision-making – as in ‘evidence-based policymaking’ – symbolic knowledge utilisation refers to more indirect functions of knowledge for policymakers, either to substantiate already-decided policy choices or to legitimise policy actors. Following Boswell’s analysis of knowledge utilisation in the UK, Germany and the EU (2009), we expect that politicisation generates more symbolic forms of knowledge utilisation.

The third and final hypothesis refers to the effect of research-policy dialogues on developments within the field of migration research itself. In the science studies literature this has been conceptualised as the ‘co-evolution’ or ‘coproduction’ of scientific knowledge (Ezrahi 1990; Jasanoff 2013). For migration studies in particular, a rich literature has evolved in this context, both supporting (Brubaker 1992; Koopmans and Statham 2000) and criticising (Vasta and Vuddamalay 2006; Bertossi 2011; Favell 2003; Thränhardt and Bommes 2010) national models of integration. This has largely been an empirical debate on whether integration policies are framed in distinct national settings, or whether scientists should take a more dynamic, multi-layered, but also more contested approach to how integration is framed. As Chaps. 4 and 5 in this volume indicate (see also Favell 2003; Bertossi 2011), this debate also speaks to how relationships between researchers and policymakers affect processes of knowledge production. This has structural implications since such a relationship may have an impact, for example, on the financing of specific research centres, but it also has cultural effects because privileging specific frames of migrant integration affects the way the general public perceives this process (see also Entzinger and Scholten 2013). Such symbioses between integration research and policymaking in Europe have often been criticised for favouring specific knowledge producers and knowledge claims while ignoring others, thus reproducing specific national models of integration (Thränhardt and Bommes 2010; Duyvendak and Scholten 2011). Turning this argument around, in this book we expect the opposite to happen when migration and integration become politicised. Developments such as the de-institutionalisation of research-policy dialogues and the internationalisation of academia challenge the container view of ‘national models of integration’, and can be expected to contribute to academic fragmentation (see also Favell 2003; Thränhardt and Bommes 2010).

In the following sections we assess the evidence for each of the three hypotheses on the basis of what has been brought forward in this book. We do so firstly by comparing the historical development of knowledge production, knowledge utilisation and research-policy dialogues around migrant integration, in different countries and at different levels. After that we look in more depth at how research-policy dialogues have evolved around three core themes in the integration debate: naturalisation, education and religion, basing our comparisons and conclusions on the much more detailed information collected in the rest of this volume. To assist the reader who is interested in more detail, we make direct reference to the relevant chapters in Part I of this book. Conversely, in order to keep the text readable, we make no references to the individual country chapters in Part II. All information, however, on specific countries as well as on the EU can be traced back to the corresponding case studies in that part of the book.

3.1 Changing Structures of Research-Policy Dialogues?

Concerning the first hypothesis on the relationship between politicisation of migrant integration and de-institutionalisation of research-policy dialogue structures, we find mixed evidence when comparing the findings from the various cases. Generally speaking, we have found evidence of changes in the institutional set-up of research-policy dialogues, rather than a clear de-institutionalisation of dialogue structures. In short, politicisation changes rather than impedes research-policy dialogues. Interestingly, in some cases we found evidence that institutionalisation followed after a period of politicisation, as in Germany and Austria. It seems as if in these cases politicisation first created a sense of urgency, which later, when the political climate had calmed down again, was translated into an institutionalisation of research-policy dialogues.

Only in the Netherlands, Italy and Denmark did we find modest direct evidence in favour of the original de-institutionalisation hypothesis. In the Netherlands the strongly institutionalised technocratic research-policy nexus that had been built around the Ethnic Minorities policy in the 1980s was indeed dismantled in the 1990s and 2000s, in a context of increasing politicisation. However, this de-institutionalisation was also spurred by developments within the research community. Furthermore, it led to a re-institutionalisation of a different type of nexus, of a more bureaucratic nature, that focused not so much on conceptual research but rather on data-driven studies, carried out by the Social and Cultural Planning Office (SCP) and Netherlands Statistics (CBS), two government agencies. In Italy attempts were made to institutionalise research-policy dialogues in the 1990s, particularly by setting up a Commission on the Integration of Migrants. However, given the politicisation of migration at that time, that commission did not have a great impact and it was dismantled. In Denmark researchers were directly involved in the genesis of policies in the 1980s and 1990s, but less so thereafter. However, as in the Dutch case, a certain re-institutionalisation of research-policy dialogues also took place in Denmark, though different in character.

Importantly, we found that in various cases, including Germany as the most prominent example, the politicisation of migrant integration may have spurred the establishment of research-policy dialogues rather than having impeded them. In contrast to other cases studied here, such as the Dutch case, the development of migration research in Germany followed a more autonomous path with very little dialogue between researchers and policymakers, especially at an institutional level. The politicisation of migrant integration at the end of the 1990s and in the early 2000s provided various opportunities for researchers to become more actively engaged in policymaking and in political debate. This led to the establishment of various ‘boundary structures’ (Entzinger and Scholten 2013), such as the Council for Migration and the Expert Council for Migration and Integration. The Austrian case, reflecting to some extent the German experience, even reveals evidence of efforts to institutionalise research-policy relations in the aftermath of politicisation. In Austria, at the end of the 2000s, the grand coalition between the People’s Party (ÖVP) and the Social Democrats (ÖVP) involved researchers and research-based commissions as well as NGOs and other stakeholders in formulating a National Action Plan for Integration (NAPI). At the same time, more informal dialogue structures emerged in Austria, outside institutional channels.

The more bureaucratic dialogue structure that emerged in the Netherlands in the late 1990s, with a preference for statistics-driven research oriented towards specific government policy priorities, also emerged in other countries. In Germany, the Federal Institute for Migration and Refugees (BAMF) performs a role in policymaking that is very similar to that of the Dutch SCP. As Boswell (2009) shows, the BAMF plays a key role by producing data that help to legitimise national policies, promote policy learning (both nationally and internationally) and monitor and identify areas for policy intervention at the national, regional and local levels. A similar process can be identified in Denmark: Bak Jørgensen (2011) describes the development of in-house research facilities at the Ministry of Refugees, Immigration and Integration Affairs, bringing together knowledge and information that is important for policy coordination.

One case that clearly seems to defy the hypothesis on politicisation and de-institutionalisation of research-policy dialogues is the EU case, which is distinctive in two ways. The first aspect concerns timing: the EU entered this policy area rather late, at a moment when politicisation of the issue was thriving in much of Europe. In this respect it resembles the Polish case, where the emergence of a policy could not lead to any form of de-institutionalisation for the simple reason that no relevant institutions had been set up yet. The Polish case differs from the EU, however, because thus far the issue has never been politicised.

A second reason why the EU case differs from all others is that the EU’s position in the multi-level governance system is completely different from that of national governments. In the absence of direct competencies in the field of migrant integration, mobilising research has proved to be one of the few strategies the EU possesses to influence policies in this domain. Geddes and Achtnich (Chap. 16) as well as Pratt (Chap. 7 of this volume; see also Geddes and Scholten 2013) show how a selective mobilisation of research has provided a tool for the soft-governance of migrant integration in a European setting. In particular this has led to a number of comparative studies of migrant integration policies aiming to facilitate horizontal policy learning between European countries. It has also led to more systematic efforts to measure integration policies so as to monitor compliance with EU policy frameworks (for example the Migrant Integration Policy Index, MIPEX). This latter strategy to promote horizontal learning by comparing policies and instruments has recently emerged at the local level as well, leading to some very specific research-policy dialogues in networks of European cities, such as Cities for Local Integration Policies, Integrating Cities and Intercultural Cities (see Chap. 6).

What stands out in the national cases in terms of dialogue structures is the central role of ad-hoc and often government-sponsored commissions at critical junctures in the policy process (see also Boswell and Hunter 2014). In Germany, Italy, the UK, and the Netherlands such commissions were put in place in the aftermath of focusing events in order to create a temporary platform for research-policy dialogues. Such commissions, however, often are highly selective in opening up to researchers and in their knowledge claims. This suggests that creating ad-hoc commissions should be seen as a political reflex in the face of immediate and intractable policy controversies rather than as an effort to engage in critical reflection based on research. Furthermore, though their public profile was often high, the policy impact of the work of these commissions was not always very direct. In fact, commissions in the UK (for example the Community Cohesion Review Team, led by Ted Cantle), the Netherlands (the Temporary Parliamentary Investigative Commission on Integration Policy, led by Stef Blok), and to a lesser degree also Germany (the Commission on Immigration, led by Rita Süssmuth) show how easily the knowledge claims selected by these commissions can lead to public controversy and to contestation of the commissions’ authority.

Furthermore, several of the cases suggest that the media play an increasingly important role in opening up dialogue structures to a broader set of actors and to the public. Dialogue structures in the rational policy model were mostly restricted to a relatively small network of scholars and policymakers. The traditional route was to communicate with policymakers directly, for instance in private without media coverage, in commissions, boundary organisations, or at conferences, informal dinners, and so on. Another route we have observed, especially in the Dutch, German, and British cases, involves communicating with policymakers indirectly, via the media. In the DIAMINT project we found numerous examples of public intellectuals who seek media attention in order to influence policymakers in this domain.

Finally, what can be observed in many countries, as well as at the EU level, is the growth of dialogue structures for purposes of policy monitoring. The German BAMF and the Dutch SCP can be considered as the most institutionalised versions of this type of dialogue structure. This also applies to MIPEX, though it monitors policy compliance rather than policy performance. A key rationale for dialogue structures of this type, which we will discuss further in the next section, is that they help identify areas for policy action and provide coordinating ministries with tools to mobilise policy intervention by other ministries and institutional bodies that possess ‘hard governance’ tools that enable them to act more directly in a specific policy area where such intervention is deemed necessary.

3.2 Towards More Symbolic Knowledge Utilisation?

Our comparative analysis provides strong support for the second hypothesis on the increasingly symbolic character of knowledge utilisation. Our analysis shows that politicisation does not impede knowledge utilisation, but rather changes its nature. All national cases show that the use of knowledge became increasingly symbolic in the 2000s. This includes forms of substantiating knowledge utilisation, where research is used to support already-decided policy choices, as well as forms of legitimising use, where research is used to boost the authority of specific policy actors.

Some of the cases, such as the British case in the 1950s and the Dutch case in the 1980s, indeed show how research initially provided a direct stimulus for policy development, thus accounting for instrumental forms of knowledge utilisation. This was also the case in Sweden, which was not included as a case study in this book (Hammar 2004). The other countries studied do not provide such clear evidence: generally we have found only incidental cases where individual academics may indeed have had an impact on policy development at key moments, as did the Vesselbo report in Denmark. This challenges the assumption in the literature that migrant integration research originally had a strong policy orientation. In fact, countries like Germany and Denmark, and to some extent also Italy and Austria, show a largely autonomous development of migration research and policymaking alongside each other, but with very little interaction.

Most of the cases examined reveal intriguing examples of symbolic knowledge utilisation. In the originally more instrumental cases of the Netherlands and the UK the use of knowledge claims clearly became more selective around the year 2000, aimed at substantiating policies formulated in the political arena. In the UK, for example, research was utilised (at least partly) to substantiate the Community Cohesion frame that emerged in politics after the ‘milltown riots’ of 2001. In the Netherlands research-policy dialogues virtually came to a halt after the political developments of the early 2000s, with government only selectively using outcomes of carefully commissioned research for purposes of policy monitoring. Similarly in Austria, there is broad consensus over the selective use of knowledge, mainly driven by the development of in-house research facilities at the Ministry of Interior. In Denmark the state supported the Academy for Immigration Studies (AMID), thus legitimising its policy position. However, it hardly ever drew on the findings and recommendations of AMID studies.

Very specific forms of legitimising knowledge utilisation were found in Germany, the Netherlands, and also at the EU level. As has been mentioned earlier, the German and Dutch governments rely rather heavily on data produced by governmental research agencies such as BAMF and SCP. In both countries the choice of indicators for which data are being collected is made in collaboration with these government agencies. As such, these institutes provide ‘mandated truths’ (Salter 1988) that are used in the complex inter-departmental and multi-level governance of migrant integration. They provide ammunition to raise problem awareness and to trigger policy interventions by other actors. This confirms what Boswell has already observed: even in an extremely politicised domain such as migrant integration policymakers are keen on upholding “the myth of instrumental use” (Boswell 2009: 249). However, the German case in particular also shows how knowledge utilisation can be symbolic and legitimising, but at the same time also instrumental, particularly when the development of more concrete policy plans and measures and their implementation are at stake. The important differentiation to be made here is that this instrumental use is found primarily in relation to secondary policy aspects, whereas on the level of more fundamental policy premises it is much more difficult to discern.

The cases of the EU and European cities (see Chap. 6) demonstrate a different dynamic. The EU case shows that precisely because of the specific setting in which it operates and the absence of concrete policy measures, the mobilisation and more symbolic use of knowledge provide important tools for the soft-governance of migrant integration policies. At first, the EU initiated comparative research to facilitate ‘horizontal policy learning’ between member states. Very often this involved comparisons at the local level, possibly because local government often plays a crucial role in migrant integration. This EU approach can be interpreted as a legitimising form of knowledge utilisation, since such comparisons and the policy convergence at which they are aimed provide a basis for EU policy intervention. The Migrant Integration Policy Index (MIPEX) can be seen as a primary example of such legitimising forms of knowledge utilisation, as it monitors the compliance of member states with EU policies, and thus further legitimises EU policy intervention.

This has evolved into more substantiating forms of knowledge utilisation. And indeed, the EU is increasingly mobilising research based on specific EU policy priorities, as formulated in the Common Basic Principles of Integration, adopted in 2004, and the Stockholm Common Integration Agenda of 2009. At the same time, one may also observe a growing role for Eurostat, the EU’s statistical agency, in data collection and in prescribing what data national statistical agencies in the member states should collect (see Chap. 3). This can be considered an indirect manner of agenda setting for the member states, comparable to the role agencies like BAMF and SCP play at a national level.

European cities have yet another position in the increasingly complex multi-level governance of migrant integration. Until 2003, their strategies and research-policy dialogues were largely determined by the question whether or not national integration policies existed at all, as illustrated by Dutch versus Swiss cities and their use of research (see Chap. 6). The very specific dialogue structures for horizontal learning that the networks of European cities have established since 2004, with strong support from the European Commission, are certainly novel. It is not clear whether the function we should attribute to these structures is primarily instrumental or symbolic. They seem to fulfil both.

In some cases, politicisation has also contributed to a growing contestation of research at large. This applies in particular to the case of Italy, traditionally characterised by a certain distrust toward social-scientific knowledge. To some extent it also applies to the Dutch case, where the credibility of migration scholars involved in policymaking in the 1980s and 1990s was openly put on the line in the 2000s; they were blamed for introducing a multiculturalist bias into Dutch policies. This reinforced a mode of ‘articulation politics’ characterised by clear political primacy, an orientation towards popular – if not populist – views and distrust, especially of research on a conceptual level. This speaks to Caponio et al. (2014) analysis of expertise being constructed not only before but also during the knowledge utilisation process. This underlines the social nature of processes of knowledge utilisation, whereby the knowledge that is potentially utilised is not ‘out there waiting’, but actively and constantly under construction and an object of on-going contestation. Many different actors – policymakers, experts, but also the media – are likely to be involved in the process, implying (partly) converging roles irrespective of their labels. As observed in our case studies, such role convergence stems from actors’ common interest in increasing the ‘epistemic authority’ of some knowledge claims and claimants while decreasing the authority of other claims and claimants (ibid.).

3.3 Knowledge Production: Beyond National Models of Integration

Our analysis of knowledge production clearly shows how migration research has vastly expanded over the past decades, not just in the ‘old’ immigration countries in Europe, but increasingly also in the ‘new’ immigration countries. Undoubtedly the ongoing nature of immigration and the continued growth of migration-driven diversity in European societies have contributed to this growth. In terms of our third hypothesis on the diversification of knowledge claims and the rise of knowledge conflicts, our comparative analysis provides partial support. We indeed observe a decreasing relevance for researchers of so-called national models of integration in the various cases that have been examined. However, this seems to be related not just to politicisation and changing research-policy dialogues in these national settings, but also to broader developments such as the growing involvement of both the EU and local authorities, and the internationalisation of the migrant integration research community.

Whereas research in Austria, Germany and Italy was fragmented even before the issue had become more politicised, the Dutch and British cases indeed show a more gradual fragmentation. Before politicisation, migration scholarship in these countries was characterised by a relative consensus within their respective national contexts, leading to what have been framed as distinct ‘national models of integration’: the Dutch ‘Ethnic Minorities’ model and the British ‘Race Relations’ model. Following politicisation (which occurred much earlier in the UK than elsewhere), these national models became fragmented and contested. Both the UK and the Netherlands also reveal many instances of knowledge conflicts amongst scholars. In other countries, for example in Germany, migration scholarship has always been more fragmented, possibly even because of the absence of an institutional relationship to policy that could have sustained a single national model of integration. In Germany distinct schools of thought can be identified that co-evolved over the past decades, in particular an ethnic-minority approach, a socio-historical approach and a rational-choice approach. However, it is fair to say that in Germany, but also in other countries, such as Austria and Italy, knowledge claims have become even more diversified meanwhile, also along disciplinary lines.

The EU case again appears to be different, due to its recent genesis: there was simply no pre-existing unity against which a possible fragmentation could have taken place. Research initiated and supported by the EU has always had a special, comparative character.

Interestingly, the fragmentation of knowledge paradigms at the national level goes hand in hand with a growing alignment of knowledge paradigms between countries. Whether an indirect effect of EU involvement in this research area or just a symptom of a broader internationalisation of the academic world – exemplified by the development of various international networks such as IMISCOE – the growth of international and comparative research generated at the EU level has no doubt had an impact on national and local research as well. Knowledge paradigms seem to be increasingly developing along the lines of academic disciplines (political science, economics, sociology, anthropology, law) and this occurs in very similar ways in the various countries that we have examined.

Finally, our analysis also shows that, contrary to what has been suggested in the literature for specific cases (Favell 2003; Scholten 2009), the development of migrant integration as a research field has not necessarily been driven by ‘co-production’ or ‘co-evolution’ with the policy context. We do indeed see evidence of research-policy coordination in some countries, especially the Netherlands and less so in the UK. By contrast, however, Germany and Denmark, but also Italy and Austria, have witnessed a largely autonomous development of research and policy. In Italy and Austria government interest in research emerged long after a (national) research field had been established. In Germany migration research evolved relatively autonomously within various institutions strongly embedded in academia. In Denmark DAMES (the Danish Institution for Migration and Ethnic Studies) spurred the development of migration research as well as its international orientation without having a significant policy orientation.

4 Research-Policy Dialogues on Naturalisation, Religion and Education

In order to obtain a more detailed view of how research-policy dialogues take shape around concrete topics in the integration debate, each of our DIAMINT case studies also provides a more in-depth analysis of three core themes: naturalisation, education and the management of religious diversity. The selected themes not only represent three core issues in integration debates throughout Europe, but they also illuminate very different dimensions of migrant integration: the politico-legal, the socio-cultural, and the socio-economic dimensions respectively.

Naturalisation, as a theme closely associated with citizenship, and with significant implications for migrants’ rights and entitlements, has been on the agenda in most countries before as well as after the politicisation of migrant integration.

The accommodation of new religious diversity in practically all cases primarily revolves around the growing salience of Islam in European society. This is one of the themes that has emerged more centrally on the agenda since the politicisation of migrant integration, though at different paces and to different degrees in the various cases.

Finally, reception policies for migrant children in primary and secondary education were chosen as a theme in the socio-economic field. The inclusion of newly-arrived migrants and their children in existing educational arrangements has also been on the agenda in all countries studied since the early days of immigration, but, as we will see, in very different ways. In what follows we will analyse what our studies of these three core themes have contributed to our understanding of dialogue structures, knowledge production and knowledge utilisation.

4.1 Dialogue Structures: Differences Between Themes and Between Countries

Our analysis reveals important differences in how research-policy dialogues are organised for the three chosen themes, both between and within the countries under study. Of these three, accommodating new religious diversity appears to be the most contested theme and it is not really surprising that the shape of research-policy dialogue structures on this issue differs significantly between all cases. In the Netherlands an informed dialogue between researchers and policymakers on religious pluriformity in general and Islam in particular has been largely absent. The same applies to Germany, although there the ‘Islam Conference’ provides opportunities for academics to exert indirect influence on policymakers. The British, Italian and Austrian cases all appear to be different again. In Austria we do see clear efforts to institutionalise a dialogue structure around this topic, via the increasingly influential ‘Dialogue Forum Islam’, in which there is also some room for knowledge production. In the UK, religion – and Islam in particular – has been central to the three government-sponsored commissions that have reported in the past decade (though with very different knowledge claims involved). However, even though these commissions have had great influence in many areas, their influence on policies relating to religion has been limited. In Italy, a special Scientific Committee has played a key role in drafting a charter on Values of Citizenship and Integration that includes a clear view on the role of religion in society.

On the theme of primary and secondary education for migrant youth, more systematic dialogue structures have emerged from our comparative analysis. Education is one of the areas where the collection of data and statistics (often government-sponsored) plays a central role in the monitoring, evaluation and mobilisation of policy initiatives. This is perhaps best exemplified by the Dutch case where data collection by the Social and Cultural Planning Office (SCP) has been an important tool for policy coordination. Those data have been used, for instance, to monitor policy performance in different sectors and, if needed, to mobilise policy intervention by the relevant Ministries.

Our analysis shows that research and statistics can also play a role beyond that of policy evaluation and monitoring. The Programme for International Student Assessment (PISA), initiated by the OECD, for example, not only revealed a relatively modest overall educational performance in certain countries by comparing these to others, but it also indicated that young people of migrant background often perform well below the national average. In some countries, particularly in Austria and Germany, the PISA studies have had a great impact on national policymaking in the area of migrant integration, setting educational performance firmly on the political agenda and spurring a range of policy initiatives.

Finally, as regards naturalisation pronounced differences have emerged between the different cases examined. In the UK, academics and other experts played a central role in the revision of naturalisation policies, for example in the Commission for the Future of Multi-Ethnic Britain, the Community Cohesion Review Team and the Advisory Board on Naturalisation and Integration. Likewise in Italy the Commission for the Integration of Immigrants has had an important effect on the reform of legislation on naturalisation. In the Netherlands the role of academics on this theme has been less institutionalised, but academics nevertheless did exert influence on societal and political debates by advocating more liberal policies. In other countries, for example in Austria, naturalisation has been treated primarily as an issue for legal experts within the government, and at best some informal dialogue has taken place externally, often also involving other actors such as NGOs.

What has also emerged from all examined cases is the role the media may play in shaping the debate between researchers, policymakers and other stakeholders on each of the three themes. Take, for instance, the key role of the media in enhancing the impact of the PISA scores in Germany and in encouraging public debate on this. We have also come across several examples of media actors who actively position themselves as participants in research-policy dialogues, which then broaden up to science-society dialogues. Certain media actors have more or less explicit political agendas, be it for opportunistic reasons or for ideological ones. In the UK, the Netherlands and Italy, for example, it is clear that some newspapers favour strict naturalisation policies and have a securitised view on Islam. We have also found examples of academics and ‘public intellectuals’ who use the media as a platform not only to communicate the outcomes of their research to a wider audience, but also to pursue political or ideological goals. This, of course, is perfectly legitimate in a democracy, but it can also be interpreted as interfering with the direct flow of ideas from science to policy (the transmission of ‘truth to power’). Seeking media access enables some critical experts to get a ‘public voice’. If it were not for the presence of open media outlets, policymakers might never hear some of these experts at all.

4.2 Conditions for Instrumental and Symbolic Knowledge Utilisation

In terms of knowledge utilisation, the DIAMINT cases reveal a complex interaction between knowledge producers, the (party-)political context and the media in all three thematic areas that we have analysed in depth. It is often hard to predict whether research outcomes will be used at all, and if so whether their use will be instrumental or symbolic. Political interests and media attention may certainly act as catalysts here, as with the ‘PISA shock’ in Germany, described earlier. There can be little doubt that the overwhelming and long-lasting media attention that PISA received convinced politicians of the need to act. Yet our more detailed analyses show that the process of knowledge utilisation is seldom clear-cut. In the UK, for example, ministers’ attitudes towards the report by the Commission on the Future of Multi-Ethnic Britain changed from positive to negative once it had become apparent that media coverage of the report was overwhelmingly hostile. Still, in the final instance the report clearly inspired new legislation in the area.

Moreover, we tentatively conclude that there is greater potential for instrumental knowledge utilisation in relation to primary and secondary education than on the two other themes we have studied. Very often, however, research in the field of education appears to be unable to influence relevant policies at the policy paradigm level. In Austria, for example, the area of education is marked by the active role of boundary actors facilitating dialogue and a request for applied science with regard to teaching German as a second language. While the benefits of intercultural education or plurilingualism at school are widely recognised by scientists, NGOs, and policymakers, the current Austrian system is predominantly run under the monolingual ‘submersive’ school model.

As in Germany, the PISA project on education has also had some impact in Austria. In the area of education, the Italian case highlights the importance of the personal commitment and interest of policymakers in a specific issue. Furthermore, prior governmental influence also emerges as an important factor: dialogue contributions directly solicited by the government seem to stand a better chance of being used instrumentally, that is to modify existing policies or legislation. This, however, does not rule out symbolic uses.

From a comparison of all DIAMINT case studies we may conclude that symbolic knowledge utilisation is most prevalent with regard to the theme of accommodating new religious diversity. In comparison to naturalisation and education, religious diversity is also the theme most frequently covered by the media and where the influence of media is strongest. This is not surprising. All over Europe religious diversity is closely associated with the growing presence of Islam. This is a very sensitive and highly politicised issue that leaves little room for serious consideration of research outcomes.

With regard to naturalisation, the Austrian case signals a lack of willingness to use the outcomes of international research, particularly if these are not in line with national ideas on policymaking. This phenomenon is by no means unique to Austria. If knowledge produced by third parties, for example by international bodies, is unwelcome, governments may argue that the organisation that commissioned the research, such as the EU or the OECD, holds no responsibility for the policy area at stake and is therefore not qualified to pronounce on national policies. It is again the PISA example that shows the limited validity of such arguments.

Finally, it should be noted once more that academics sometimes actively seek to engage in research-policy dialogues via the media, thus following their own ‘symbolic’ interests, such as the wish to reproduce their expert status or to secure the relevance of their academic work or their specific organisation. This not only occurs at the level of individual researchers, but also at an institutional level. In the UK and the Netherlands in particular we found universities and research funders to be quite keen on generating media attention for their projects. Some researchers experience this pressure as an annoying side-effect of their work, whereas other scientists have developed close ties with journalists and co-create news stories with their media contacts on a regular basis.

4.3 Knowledge Production: Knowledge Conflicts

Moving to knowledge production, the various cases reveal clear conflicts between researchers on how to classify the three themes selected for our in-depth analysis. Not all stakeholders necessarily share our analytical categorisation of ‘naturalisation’, ‘religion’, and ‘education’ as politico-legal, socio-cultural and socio-economic issues. Differences exist among and between academics, politicians and others in the way they problematise and frame some major issues. Nowadays, for example, issues related to Islam are more often framed as security issues in both the political arena and the media.

In the Netherlands and Germany, for instance, the so-called ‘radicalisation’ of Muslim youth has been a key concern in the past decade. As a consequence, funding for research on this particular topic has been made available rather generously. The effect of this may be that in the perception of policymakers and the general public ‘combating radicalisation’ runs the risk of becoming a proxy to ‘accommodating religious diversity’. In fact, in the Netherlands one of the earliest research projects into Islamic schools was carried out by the secret service AIVD, as reported by the Blok Committee (2004). It was also found in several cases that the public debate with regard to Islam has been quite different in tone and substance from the debate on other ‘new’ religions ‘imported’ by newcomers.

In the UK the Commission on the Future of Multi-Ethnic Britain noted that “[a]ggressive hostility to Islam is expressed in ways unthinkable in relation to other beliefs” (CMEB 2000: 10). As time went by, however, this Commission’s proposal to improve the legal protection against discrimination of Muslims was adopted (ibid.). In other words, the Commission framed the issue of ‘new religious pluriformity’ not as a problem of socio-cultural ‘self-separation’ of immigrant groups – or even as ‘backwardness’ of Islam, as Pim Fortuyn did in the Netherlands around the same time – but as a challenge for the receiving society to uphold its own pre-existing liberal democratic principles, which already protected Jews and Sikhs.

In the dialogues on naturalisation, social scientists and legal experts often adopted more liberal positions than policymakers did. In the Netherlands and Germany, for example, academics have advocated the acceptance of dual citizenship and presented naturalisation as an instrument in the integration process rather than as its end stage. Conversely, public opinion and media analysis reveal that having a single passport is perceived as a proxy for loyalty to the ‘host nation’; thus securitisation processes are at work here as well. Several of our case studies also indicate that concrete empirical evidence on the relationship between naturalisation and integration (measured, for example, as labour market participation) is very scarce (see OECD 2011). As a consequence, the contribution of scientists to the naturalisation debate often focuses on philosophical and normative arguments rather than empirical arguments, as, for example, in Germany.

The main fault line in the domain of education centres on the question whether education should be approached as a socio-economic issue, which calls for emphasising equal opportunities for migrant children in the national educational system, or rather as a cultural issue. In the latter frame, for instance, reducing educational segregation is seen as an end in itself irrespective of the effects of such segregation on educational achievements of individual pupils. Another example of cultural framing came when Dutch and Austrian politicians denounced mother tongue teaching to migrant children on the grounds of principle, irrespective of the potentially positive long-term effects that teaching other languages may have on the educational careers of migrant children and on their integration in society.

The framing of issues and differences in policy objectives and priorities have consequences for the kind of knowledge required. From the policymakers’ perspective the production of such knowledge is not necessarily a purely academic affair. We have observed repeatedly that, as the area of migrant integration has become more politicised, governments become more selective in the types of research they use and support, while they also prefer to produce their own data more often. In specific cases this may lead to a dissension of opinions when it comes to generating and using comparable data on migrant integration at the European level. Not surprisingly, the EU’s efforts to develop common indicators for integration have been only partially successful so far (see Chap. 3).

5 Conclusions

Our analysis has revealed profound changes in the dialogue structures associated with the research-policy nexus in the domain of migrant integration, rather than a clear de-institutionalisation of these dialogues, as we had initially expected. On the one hand, we have found that dialogue structures have become more ad-hoc, often established in response to distinct political events or to specific problems. A striking similarity between the countries involved in this study is the establishment of ad-hoc government commissions to channel research-policy dialogues at critical junctures in policy development. On the other hand, we have also found that politicisation has not thwarted all efforts to develop more institutionalised dialogue structures between producers and users of knowledge. ‘Going technical’ or mobilising specific types of research should not always be seen as a tactic of depoliticisation, but can also represent a strategy of ‘politics by different means’.

Additionally, we found that in all cases studied here research-policy dialogues have gradually become more open. Though it is difficult to establish a clear relationship with politicisation here, we can speak of a gradual evolution from ‘research-policy dialogues’ to broader ‘science-society dialogues’. In all countries analysed several new actors have emerged as participants in these science-society dialogues. In the British case this can be illustrated by comparing the composition of three independent commissions created to advice the government at different moments in the past decade. The membership of these commissions not only consisted of academics, but also included expert practitioners in areas such as law, health, local government, education, and journalism. Interestingly, the most influential of these commissions (the Community Cohesion Review Team led by Ted Cantle) left out academics altogether. In the Netherlands, the role of ‘public intellectuals’ in research-policy dialogues has increased noticeably; relevant names here are Paul Scheffer and Jaap Dronkers. From the Italian case study it becomes apparent how important (primarily faith-based) NGOs can be in providing knowledge to policymakers. The Austrian case study documented a central role for social partners alongside the very dominant Ministry of Interior. In Germany it was found that civil society initiatives, such as the Academies, were very open to diverse actors, playing an important stabilising role in research-policy dialogues on nationality legislation and on Islam. In all cases the DIAMINT researchers showed that the media should be conceptualised both as a platform for research-policy dialogues and as an important participant in such dialogues.

In the rational model of governance knowledge utilisation has traditionally been assumed as being direct and instrumental. Our analysis provides clear evidence that more symbolic forms of knowledge utilisation prevail in practically all cases. Knowledge is being used primarily not in an instrumental manner, but rather to legitimise government institutions or to substantiate government policies. This required, amongst other things, the mobilisation of specific data to be used as soft-governance tools for policy coordination (for instance through institutions such as BAMF, SCP, MIPEX) as well as the establishment of expert commissions to substantiate policy frames (such as the ‘Community Cohesion’ frame in the UK) or to justify abandoning existing frames (as the ‘Germany is not an immigration country’ frame in Germany). At the same time we can observe that, particularly in the Netherlands and Denmark – both older immigration countries that witnessed a substantial shift in their dominant policy approaches around 2000 – research-policy dialogues on the more conceptual level of policy framing have virtually come to a halt.

Furthermore, most cases provide evidence of the growing role of the media in processes of knowledge utilisation. Academics wishing to have a policy impact can basically opt for two routes. The first, more traditional route is to communicate with policymakers directly. In practice this mostly happens in private, without media coverage, in so-called ‘boundary organisations’ like advisory bodies, special commissions and think-tanks, and it also happens at conferences and informal workshops. A second route is to communicate with policymakers indirectly via the media. In this project we have found numerous examples of academics and public intellectuals who have developed media strategies to influence migrant integration policymaking. There is some evidence (particularly from the British and Dutch cases) that to a certain extent these two routes can be considered mutually exclusive. This is because communicating via the media has a ‘cost’: put simply, scientists who choose the public route tend to take more critical positions and this decreases the likelihood that they also interact with policymakers directly in more private networks. There is a link between the tone and content of researchers’ messages, their media presence, and the usefulness of their expertise as perceived by policymakers.

Finally, our analysis has revealed a sharp increase in both the quantity and diversity of knowledge production and knowledge dissemination over the past decades in all cases examined. This makes it more complicated than in the past to identify or to construct a distinct research-policy nexus. Nowadays, many centres for knowledge production exist. Some are more academic and mono-disciplinary, while others are more akin to boundary organisations, deriving expertise from different disciplines involved in this area. Some use primarily quantitative methods, while others specialise in qualitative research. Some focus on comparative research, while others are more nationally or even locally oriented. It is beyond doubt that this has also facilitated a more selective use of knowledge claims within specific policy settings. It also signals, however, the maturing of migrant integration as a research area, in which room has developed both for more policy-oriented and for more theory-oriented schools.

Interestingly, this expansion and diversification appears to have taken place in the ‘old’ as well as in the ‘new’ immigration countries. This is not just an endogenous effect of all countries being increasingly confronted with migration and diversity, but it has also been provoked by the internationalisation of academic research as such. The European Union has played a dominant role in this process by facilitating the comparative research needed to support its growing involvement in this policy area. In addition, the research community itself has also been active in reaching out across national borders. IMISCOE, for example, one of Europe’s largest research networks in the domain of migration and integration has fulfilled an important role in incorporating Central and East European scholars into the international research community.

Furthermore, our analysis provides evidence of how knowledge production and knowledge utilisation may influence one another. The changing nature of dialogue structures and new forms of knowledge utilisation have eroded nationally-oriented framings and models of migrant integration in those relatively few cases where these had come into existence. This evolution of dialogue structures and knowledge utilisation has provoked a fragmentation of migrant integration research in terms of numbers of institutes involved and it has also led to a diversification of knowledge paradigms. As an academic field of study migrant integration has come of age.