A relational approach to local immigrant policy-making: collaboration with immigrant advocacy bodies in French and German cities

ABSTRACT The role of immigrant advocacy bodies in collaborative policy–making in cities is so far insufficiently researched. This article investigates the ties between relevant urban actors and immigrant advocacy bodies in cities in two Western European countries. We draw on an original survey in forty French and German cities as well as fieldwork in one French and one German city to analyze whether urban actors from a variety of policy sectors and domains of society cooperate with immigrant councils and immigrant associations, and which factors explain such collaboration. Counter to the existing literature on the role of intermediaries between municipalities and immigrant populations, we find a widespread existence of ties with immigrant advocacy bodies. However, such ties are not mainstreamed. Instead, collaboration is most present among actors in charge of immigrant affairs, and when actors meet in policy fora that allow interaction between urban actors and immigrant advocacy bodies.


Introduction
Representation of immigrant interests has existed for many decades in European cities, but often immigrant advocacy bodies were not recognized as partners by urban administrative, political, economic and civil society actors (Però and Solomos 2010;Schrover and Vermeulen 2005;Thränhardt 2013). In the meantime, top-down policy-making has been complemented by collaborative forms of policy-making, also referred to as a trend towards governance (Klijn 2008). New networks of stakeholders have been created and new practices of collaborative policy-making have emerged in cities (Aarsaether, Nyseth, and Bjorna 2011;Ansell and Gash 2008;Barnes, Newman, and Sullivan 2006). These can foster the participation of immigrant advocacy bodies and the inclusion of immigrant perspectives in policy-making. Has the distance between urban actors and immigrant advocacy bodies decreased as a result of these developments? Are urban actors considering immigrant advocacy bodies as important collaborators in policy-making today? Some research has investigated the ties of collaboration between a variety of urban actors and immigrant advocacy bodies in the context of policymaking (Caponio 2005;Pilati 2012;Triviño-Salazar 2018). Research has shown, also for Germany (Halm 2015;Thränhardt 2013) and France (Downing 2015(Downing , 2016, that such ties are fragile and often dependent upon public funding, which implies the risk of co-optation. Yet there is still much to be done to fully capture the patterns of collaboration between urban actors and immigrant advocacy bodies in light of a trend of collaborative governance. We assume that in a context where networks and collaboration are emphasized as key means of policy-making, ties with immigrant advocacy bodies are ascribed more importance, and that those ties allow to gather important information, generate agreement, and legitimize policies within a diverse polity.
In this article, we analyze the ties between relevant urban actors and immigrant advocacy bodies in processes of policy-making in big French and German cities. The main questions guiding the research are whether and to what extent urban actors collaborate with immigrant councils and immigrant associations, which factors enhance such collaboration, and why.
We draw here on data yielded from a survey with a range of urban actors in 40 large cities in Germany and France, sampled based on their population size, complemented by insights from qualitative case studies in one French and one German city. The focus of the article is to aggregate findings from 40 large cities in both countries but not to perform a country comparison, because more systematic comparison of cities and urban governance of migration and diversity is needed (Martinez-Ariño 2018), of which actor collaboration is an important element.
The article is structured as follows. We first briefly review the relevant literature on the presence of immigrants and of ties between state and non-state actors in policy-making. We then present the data and methods of our study in German and French cities. The next section discusses our findings, divided into three parts: (a) the descriptive results concerning collaborations with immigrant advocacy bodies; (b) the regression analysis to elucidate factors that increase the likelihood of such collaborations; and (c) an interpretation of the mechanisms underlying such effects based on qualitative fieldwork in two cities. We conclude elaborating on the contributions of our study and suggesting avenues for future research.

Local immigrant political participation and governance networks
Immigrant political participation has been a matter of academic interest for some time. This has been approached in different ways, focusing on the representation of immigrants in parties and parliaments (Garbaye 2005;Schönwälder, Sinanoglu, and Volkert 2013), on immigrant organizations and collective forms of mobilization (Però and Solomos 2010), and on fora of deliberation created by governments to involve immigrants and their organizations in decision-making (Bausch 2011;Cerrato Debenedetti 2010;Takle 2015). Scholarship has focused on whether immigrant organizations can represent immigrant populations vis-à-vis policy-makers and whether their existence contributes to the integration of immigrants (Schrover and Vermeulen 2005). However, there is much less research on how different urban actors reach out to immigrants and their organizations.
An established strand within that literature argued that immigrant organizations compensate for lack of political opportunities and increase opportunities for participation in formal and informal politics (Eggert and Pilati 2014;Fennema and Tillie 2004). Other research has analyzed the direct interrelations between municipal actors and immigrant populations in policymaking (De Graauw and Bloemraad 2017;Ireland 2016;Nguyen Long 2015;Nicholls and Uitermark 2016;Però 2007;Uitermark 2012;Uitermark and Duyvendak 2008) stressing the role of intermediaries, such as large welfare organizations and political parties, implying that immigrant advocacy bodies themselves are rarely direct partners of the local administration (Caponio 2005;Pilati 2012;Triviño-Salazar 2018) and when they are, the relationship is fragile (Downing 2016). Counter to such findings and because of a trend of collaborative governance, we expect to find direct links and interactions between urban actors and immigrant advocacy bodies.
Looking towards institutionalized governance structures, some literature has posited that an integration policy field has emerged in cities, where a defined set of bureaucratic actors (often in interaction with collective immigrant actors) are concerned with deliberating, advocating and selecting different courses of action in immigrant policy-making (Bousetta 1997, 221). Inversely, another strand of the literature posits that an integration policy field has dissolved and actors from different policy sectors, such as education, housing and employment, take responsibility for integration. Authors refer to this as mainstreaming (Scholten and van Breugel 2018). Following the first line of argumentation, we expect that collaboration happens primarily within an integration policy field, i.e. that there is close collaboration between those urban actors with the mandate of working on immigrant incorporation and immigrant councils and immigrant associations. A different, more extensive, pattern of collaboration would exist if actors outside of the integration policy field, e.g. actors working on education or urban planning, as well as economic actors and other civil society actors, also collaborated with immigrant advocacy bodies in the city. Furthermore, we assume that collaboration is not only determined by the character of the policy field, but also by the ways in which actors more generally conduct local politics.
Political scientists commonly claim that interrelationships of state and nonstate actors with the goal of promoting a certain service or policy, denoted as a trend of governance, have become more important (Jessop 1995;Klijn 2008). Governance is also considered as increasingly present in the design and implementation of public policies in cities (Aarsaether, Nyseth, and Bjorna 2011;Nyseth and Ringholm 2008), including French (Pinson 2010) and German cities (Bogumil and Holtkamp 2004;Sack 2012). Against this background, building networks is a crucial precondition for participation in politics (Campbell 2013;Mcclurg 2003;Teorell 2003). In the qualitative analysis of this article, we focus particularly on collaborative governance, which "brings public and private stakeholders together in collective forums with public agencies to engage in consensus-oriented decision making" (Ansell and Gash 2008, 543). More specifically, we offer empirical insights into the modus operandi of fora of collaboration and their effects on the establishment of ties between urban actors.
Proposing a relational approach to urban immigrant policy-making that systematically generates and analyzes empirical data on actors' relationships, we focus on the ties of collaboration between important urban actors and immigrant advocacy bodies in German and French cities. The former include political, administrative, economic and civil society actors, 1 whereas the latter include immigrant councils and immigrant associations. 2 While immigrant councils are more common in German cities, they also exist in French cities. Other actors may draw on immigrant advocacy bodies when they want to have the interests of immigrants considered in policy-making. Our aim is to find out whether urban actors consider these bodies as collaboration partners in urban policy-making, which factors increase the likelihood of such collaboration, and why.
In their comparative study of Berlin, Amsterdam, New York City and San Francisco, De Graauw and Vermeulen (2016) consider the interaction of three local factors as determining the implementation of local immigrant integration policies, namely the existence of left-leaning governments, a high proportion of immigrants in the city electorate and decision-making structures, and the existence of a civil society that represents the interests of immigrants in local policy-making. Drawing on this literature, we expect that the likelihood of establishing an intense collaboration with an immigrant advocacy body will be higher in bigger cities where the share of immigrants is generally higher, in cities with a higher share of foreign-born inhabitants, and in cities leaning towards the left side of the political spectrum. We also expect that the likelihood of building collaborations with immigrant advocacy bodies will be higher for actors involved in local policy fora, where regular contact takes place.

Research design and data
We use an explanatory sequential mixed methods design (Creswell 2013), where ethnographic fieldwork is used to interpret the findings from a survey. Both sets of data stem from the Cities and the challenge of diversity project 3 conducted at the Max Planck Institute for the Study of Religious and Ethnic Diversity. 4 A survey, conducted between April and July 2015 in German cities and from September 2015 to March 2016 in French cities, targeted relevant urban actors in 20 large cities in Germany and 20 large cities in France. 5 Table A1 in the appendix shows the cities included and their main characteristics. In this article, we do not compare the two countries. Instead, we are interested in a systematic comparison of the existence of collaboration among urban actors across big European cities. Including cities from these two countries allows us to consider general trends in similar, yet contrasting Western European contexts, as these two countries have both a long tradition of immigration but different approaches to immigrant policy making.
The survey asked urban composite actors (Scharpf 1997, 114) about their collaborations with other actors in relation to local policy-making. We posed the question: Politics in cities is nowadays often conducted in networks. If you think about the last twelve months, with whom, and how intensively, have you collaborated in the context of your professional work or working for your organization? We refer here to collaborations in the context of politics in your city.
The survey question listed types of urban actors, including local political parties, different departments of the local administration, trade unions, economic actors, and actors representing minority interests (migrant organizations, youth organizations and disabled organizations, among others) (see appendix 1).
Our research examines ties of a range of urban actors with local immigrant councils and immigrant associations as most likely collaboration partners for other urban actors when they seek to involve the perspectives of immigrants in policy-making. Immigrant councils are consultation bodies created by city councils and composed of immigrant representatives and, in some cities, members of the city council and the administration. They exist in large German and, to a lesser extent in French cities, where they are less consolidated. They advise local policy-makers and administrations and channel information to immigrant communities (Bausch 2011;Flamant 2016). There is an ongoing debate about how independent such councils are from local administrations or whether they co-opt immigrants (Schiller 2018). Yet, we ascertain that these bodies provide an important platform for immigrants to have a say in local policy-making, as third-country nationals do not enjoy local voting rights in France and Germany. Immigrant associationsusually composed mostly of foreign-born membersare another important platform for immigrants to organize and have their interests represented, and they can be found in all large cities. Immigrant associations in Germany have been supported and strengthened over the years through public funding, but also by being recognized as experts and cooperation partners (Halm 2015). In France, the creation of organizations based on ethnicity has been and still is difficult, because they may be deemed communitarianist and anti-universalist (Montague 2013). However, municipalities subsidize a wide variety of associations, including those formed by immigrants.
The Cities and the challenge of diversity survey targetted actors likely to intervene in policymaking on the integration of immigrants and other minorities. 6 The response rate of the survey (see distribution of respondents in Table 1) is 45 per cent for German cities (n = 445) and 21 per cent for French cities (n = 249). 7 The wide scope of urban actor selection is, to the best of our knowledge, the first of its kind (see Baglioni and Giugni 2014 for a similar selection procedure).

Analysis
The analysis of the survey data on collaborations was performed in two stages that correspond to the first two result sections of this article. First, we pooled the data collected across the cities to show descriptive results on the ties of collaboration with immigrant advocacy bodies. We aim to identify the types of actors that are more closely connected with immigrant advocacy bodies, and to analyze whether ties of intense collaboration happen within an integration policy field or whether immigrant integration is addressed in a variety of policy sectors, which could be the result of mainstreaming immigrant integration. Second, a multilevel logistic regression analysis was performed that inquired about the factors that increase the likelihood of nonimmigrant urban actors collaborating intensively with immigrant advocacy bodies. The dependent variable is defined as "intense collaboration with an immigrant advocacy body". The independent variables are (1) type of composite actor, (2) participation in a local policy forum, (3) characteristics of the cities and (4) country.
By "type of composite actors", we distinguish urban actors positioned in different policy sectors and state or non-state actors. This typology helps us analyze a possible correlation between type of actor who responded to the survey and this actor's likelihood of building ties with immigrant advocacy bodies. The survey also asked respondents about "participation in a local policy forum" within the previous year, and we categorized their answers into types of policy fora, namely policy fora focused on migration and diversity issues or policy fora focused on other issues. 8 We assume that such fora are spaces of repeated and routinized interactions that can generate interorganizational ties (Levine 2013) and may facilitate collaboration with immigrant advocacy bodies in the short, medium and long term. Regarding the characteristics of the cities, we investigated the influence of (a) population size, (b) the percentage of foreign-born inhabitants, and (c) the predominant political orientation (percentage of left-wing voters in the last local election).
The multilevel model takes into account the hierarchical nature of nested data (see Snijders and Bosker 1999). In our case, it takes into consideration that composite actors of the same city might not be independent of each other, but share an urban setting that might explain their behaviour towards collaborating with immigrant advocacy bodies. By using multi-level regressions, we can assess the effects of actor variables controlling for their belonging to a particular city and city-related factors. As commonly accepted in these types of studies, the models generated in the multilevel regression analysis and presented in the second empirical section include the independent variables as fixed effects and city variance as a random effect. They have been fitted using a maximum likelihood approach (Laplace approximation) using the glmer function (from lme4 package) in R statistical software.
The analysis of the qualitative data was carried out inductively from 2015 to 2016, following a Grounded Theory methodology that starts from the data to build concepts and theory instead of using pre-conceived theoretical concepts to analyse the data (Strauss and Corbin 1990). Mannheim and Rennes were chosen because these cities had experimented in the previous years with new forms of involvement of immigrant residents, installing new fora for interaction between the municipality and immigrant residents. Participant observations in fora meetings (three in Rennes, two in Mannheim) are complemented by interviews with local officials (two in Rennes, two in Mannheim) and an analysis of publications and websites pertaining to these fora. For this article, we used the codes that capture the links between actors and activities that sought to foster collaboration within these fora. Through this, we analyze how the functioning of urban policy fora may enhance the formation of ties of collaboration.

Empirical results
The two initial questions we sought to answer with our quantitative analysis are whether and to what extent urban actors collaborate with immigrant advocacy bodies, and which factors enhance that collaboration.

Who collaborates with immigrant advocacy bodies?
This section presents descriptive results on the extent to which a range of composite urban actors build intense ties with immigrant advocacy bodies. As Table 2 shows, such collaborations are strongly circumscribed within a field of actors that focuses on immigrant issues: the branches of the administration concerned with immigrants' integration as well as immigrant advocacy bodies themselves establish more intense collaborations with immigrant advocacy actors. This stands in contrast to private and state economic actors, such as business actors and administrative actors working on economic development and labour market issues, who develop fewer intense ties with immigrant advocacy bodies. Other branches of the municipal administration related to its internal organization (i.e. human resources and citizens affairs) and to urban planning, as well as political actors and trade unions also establish fewer collaborations with immigrant advocacy bodies than administrative units responsible for integration.
The findings show that ties are mostly built among actors within a local integration policy field. Less intense collaboration is found for local public agencies working on economic and urban planning policy. One possible interpretation is that economic and urban planning departments do not consider it necessary to have intensive links with immigrant advocacy bodies, and that they consider immigrant issues or claims as primarily dealt with by integration officials. This finding challenges the assumption of a widespread mainstreaming of responsibilities for immigrant integration.

Which factors enhance this collaboration?
We now turn to analyzing the factors that enhance intense collaborations of non-immigrant actors with immigrant advocacy advocacy bodies.
Following the null model with no independent variables, several sets of models were generated, including both independent models with the four types of independent variables (actor type, actor's participation in local fora, city characteristics and country) separately and a full model that combines all variables. 9 The results shown in Table 3 confirm previous results regarding the distance between public and private economic actors and immigrant advocacy  Table 3. Regression coefficients and standard errors (between brackets) of binary multilevel regression models predicting intense collaborations of nonimmigrant actors with an immigrant advocacy body by actor type (m1), actor's participation in local policy for a (m2), city characteristics (m3), country (m4) and full model with all variables at the same time (m5). Signif. codes: "***" 0,001 "**" 0,01 "*" 0,05 "+" 0,1.
bodies, as well as the circumscription of collaboration ties mostly within the immigrant policy field. They also indicate that the participation in local policy fora increases the likelihood that a composite actor has an intense collaboration with an immigrant advocacy body. Results show that participation in both what we called "migration-related fora" and participation in fora concerned with other issues or policy sectors, such as social affairs and urban development, increases the likelihood to collaborate intensively with immigrant advocacy bodies. This suggests that such fora in general foster the creation of ties of collaboration between local actors and immigrant advocacy bodies. However, participation in migration-related fora has the highest impact on the existence of such links, suggesting that dedicated fora in the integration policy field are relevant spheres for creating collaboration between urban actors and immigrant advocacy bodies. The third model presents the effects of city-level variables. It shows that, despite the small relevance of the city level in itselfindicated by the low intraclass correlation coefficient (ICC), 10 the share of foreign-born people and the political orientation of the electorate are factors at the city level that enhance the propensity of non-immigrant actors to collaborate with immigrant advocacy bodies. The fourth model indicates that composite actors in German cities are more likely to build intense collaborations with immigrant advocacy bodies than in French cities. However, in the full model only the type of composite actor and the participation of actors in local policy fora present significant effects, which means that variation in the existence of intense collaboration with immigrant advocacy bodies is not explained by respondents' embeddedness in different cities and countries, all other factors being equal, but by the type of actor and particularly by the participation of actors in local policy fora.
Understanding the mechanisms that enhance ties of collaboration: qualitative findings One of the most interesting findings of our quantitative analysis is that participating in policy fora increases the likelihood that an urban actor intensively collaborates with immigrant advocacy bodies. But how do urban actors actually foster collaboration in such fora? Is it simply presence in the meetings that promotes the establishment of ties (Levine 2013) or are there other mechanisms at play? In this last section we provide an exploratory interpretation based on fieldwork in the "Rennes au Pluriel" and Mannheim "Bündnis für Vielfalt" policy fora.
"Rennes au Pluriel" (Rennes in Plural) is a municipal consultative body set up by the municipality of Rennes in December 2015 as part of a broader city plan to reinforce local democracy (La Fabrique Citoyenne). It is formed by 60 individual citizens and representatives of civil-society associations selected through a public call for candidatures. Its official objectives are: (1) to encourage citizens to express themselves on issues of the promotion of equality and the fight against racism and discrimination; (2) to be a space of reflection, monitoring and exchange of experiences and points of view; and (3) to coconstruct and monitor the public policy against discriminations. One of its main tasks is to organize the two-week "Rennes au Pluriel" festival.
In the German city of Mannheim, the "Alliance for Diversity" (Bündnis für Vielfalt) is an instrument to foster collaboration between different urban actors working on different dimensions of diversity, such as gender, migration, sexual orientation, age and disability. Instigated by a funding programme of the German Ministry for Families, the Elderly, Women and Youth and cofunded by the City of Mannheim, it exists since 2016. Its main activities entail networking meetings to allow participants to build up new ties of collaboration with other actors, a funding scheme for projects where actors work together across different dimensions of diversity, the organization of a one week festival ("einander.Aktionstage"), where these different projects are showcased to the whole city, and the prevention of conflicts between different population groups by signing and ratifying the "Mannheim declaration for living together in diversity". 11 By observing the activities and interactions in the two fora and drawing on interviews with officials who coordinate them, we identified four collaboration-supporting mechanisms: Participation fora are sites where municipal administrations actively try to foster a shared perspective and sense of belonging that cross-cuts specific identifications of its members (Barnes, Newman, and Sullivan 2006). Commonalities between members can increase levels of trust, and, ultimately, lead to new and sustained ties of collaboration (Klijn, Edelenbos, and Steijn 2010). In our case studies, the two fora promoted a perspective on diversity that considers several dimensions of difference as interconnected (such as gender, sexual identity, immigrant background, etc.) and cooperation across different actor-types. In the case of "Rennes au Pluriel", for example, the responsible city councilor presented the diversity of the committee members as a source of wealth, in a context where everyone should find his or her own sense and place. If such a goal is met, we argue, it is more likely that actors interact with each other and trust each other more, and potentially establish new collaborations.
Furthermore, collaborative policy fora may also work towards creating a shared vocabulary and culture that allow communication and understanding between members and ultimately facilitate collaboration. Two of the first meetings of "Rennes au Pluriel" were aimed at ensuring that all its members "had a common culture in mind" about discrimination and diversity (interview with the committee's coordinator). This was done through trainings offered to the members by external experts. Similarly, the "Mannheim declaration for living together in diversity" was meant as "a way to reach agreement ("Verständigung"), a Mannheim language, on how things are happening here. And this process of reaching an agreement, bringing groups into exchange with each other, stood at the outset of the 'Alliance for Diversity'" (Interview with a city official).
Fora provide a space to develop projects and means of action with each other that steer the action of their members in a common direction (Barnes, Newman, and Sullivan 2006). In the case of Mannheim, shared projects were fostered by making funding for projects conditional on the cooperation of actors across different dimensions of diversity. One of the resulting projects was a collaboration between the city's DITIB mosque with the city's shelter for battered women. Together, they initiated a women's group within the mosque that provides consultation to women suffering domestic violence. The project fit the goals of the Alliance, as it cross-cuts the dimensions of religion and gender and thereby "cooperation partners open up to diversity and strengthen their competences in dealing with diversity" (Interview with a city official). "Rennes au Pluriel" set as its main objective for its first working period (2015-2017) the revision and monitoring of the "Municipal plan of the fight against discrimination". Members were also asked to participate in the evaluation of projects submitted to the participatory budjet of the city. By taking part in these activities, committee members were expected to develop a sense of co-responsibility, expected to eventually lead to collaborations beyond the time frame and tasks of the forum.
Participatory policy fora are also a means to foster capacities and skills (Geissel 2009;Michels and De Graaf 2010;Murray, Tshabangu, and Erlank 2010), and hence serve as what has been captured as "schools of democracy" in the literature (Takle 2014). In the case of "Rennes au Pluriel", the coordinator invited an expert on discrimination in France, who provided participants with basic knowledge on the French legal framework on discrimination. This was meant to train them in recognizing discrimination, and to transfer this knowledge to their respective associations. In Mannheim, the Alliance sought to foster recognition among its members of each other's worth. They did so by enlisting members of the forum as selection committee for project funding. According to the coordinator of the Alliance, the learning effect was that committee members after some time moved away from supporting projects that would solely benefit their own community to supporting projects that cross-cut different communities. Professionalization through the fostering of capacities and skills can make different actors operate on eye-level and thereby facilitate collaboration. Identifying these four mechanisms of creating a shared perspective and sense of belonging, a shared vocabulary and culture, shared goals and projects, and capacities and skills helps us understand the conducive effect of participation in fora for collaboration between participating actors.

Conclusion
In this article, we have contributed to an evolving field of research on local immigrant policy-making and immigrant political participation by focusing on relations of collaboration between municipal actors and immigrant advocacy bodies. By combining a large survey as well as in-depth fieldwork, we systematically study relations in urban immigrant policy-making and gain valuable insights into the ties that a range of urban actors build with immigrant advocacy bodies. Our findings show that a variety of actors in French and German cities reach out and intensively collaborate with immigrant councils and associations representing immigrants' rights. This finding reflects larger trends towards increased recognition of immigrant advocacy bodies as well as towards more collaborative forms of policy-making.
Yet, not all types of urban actors have strong linkages with these bodies. Actors from policy sectors such as economic development and urban and territorial planning to date collaborate to a lesser extent with immigrant advocacy bodies than administrative actors responsible for immigrant integration, though some collaboration exists. The likelihood of an intensive collaboration with immigrant advocacy bodies depends on the type of actor that collaborates, as our regression analysis shows. Being a city unit responsible for integration increases the likelihood of establishing intense collaboration with an immigrant advocacy body. Most intense collaborations with immigrant advocacy bodies remain within the circle of "usual suspects" of actors engaged in immigrant policy-making, confirming our expectation that an integration policy field is a relevant context for collaboration. Possibly, this is due to the lack of mainstreaming and a perceived lack of responsibility for policy issues relating to immigrant integration by actors outside of the integration policy field, such as economic actors or urban planners. At the same time, we altogether find that collaboration with immigrant advocacy bodies has become, to varying degrees, a common element of policymaking in German and French cities. Immigrant advocacy bodies are no longer considered mere recipients of subsidies and support, but have acquired the status of collaboration partners for state and non-state actors.
Furthermore, the regression analysis has shown that it is not only the type of actor that matters. An actor's participation in a local policy forum also increases the likelihood of establishing intense collaboration with an immigrant advocacy body. Non-immigrant actors in cities with a larger share of immigrants and leaning towards the left side of the political spectrum build more intense collaborations with immigrant advocacy bodies because they participate more in local policy bodies, particularly in migration-related fora. This indicates that institutional arrangements at the city and country level matter for the establishment of collaborations.
Based on qualitative findings we outline why such collaborative fora can serve as motors for the build-up of ties of collaboration with immigrant advocacy bodies. These fora are not simply spaces that create repeated interaction. Municipal administrations also seek to change the quality of the contact by generating a shared perspective and sense of belonging, a common culture and language, and shared projects. Moreover, by training their members and professionalizing them, they seek to build up capacities and skills necessary to facilitate cooperation, and which certify immigrant advocacy bodies as respectable partners for other urban actors.
All four mechanisms combined, we argue, are likely to increase mutual knowledge and trust (Klijn, Edelenbos, and Steijn 2010). They foster the creation of "communities of practice" (Adler 2008), defined as likeminded groups of practitioners who are informally as well as contextually bound by a shared interest in learning and applying a common practice (Adler and Pouliot 2011;Wenger 1998). More research is needed to determine whether relations between urban actors in the immigrant policy-field and collaboration that stems from interactions in policy-fora change the design and implementation of policies. We have shown that the collaboration necessary for joint policy-making exists. Notes 1. In our survey we targeted composite actors ("komplexe Akteure"). Following a terminology suggested by Scharpf (1997, 114), the term "composite actors" covers corporate (korporative) actors that have some degree of formal organization and collective actors, that is, looser umbrella structures or social movements. Criteria for the selection of actors were that they were (1) involved in local politics and (2) likely to intervene in diversity-relevant fields. We excluded organizations that may have local offices but mainly formulate claims at the national level (e.g. some environmental or human rights associations). We also excluded organizations that may inform local policy decision-making, but are organized on the regional or national level. We only included actors with a minimum level of organization, i.e. when they had an office and identifiable representatives. We thus do not capture shorter-term forms of social mobilization. We also aimed to only include types of actors that could be identified across the different cities at least within the same country. We accepted that we may thus not include possibly relevant actors in specific cities (see also Moutselos et al 2017). 2. We use the term "immigrant council" to refer to all kinds of councils set up locally to address immigrant affairs in local policy-making. They are institutionalized in different ways across cities and names differ (e.g. "immigrant council", "foreigners council" or "local representation of foreigner's interests"). As targets of the survey analyzed here, we focused on organisations that frame themselves as "immigrant associations" and/or whose main aim is to advance the interests and rights of immigrants, which may exclude some initiatives by immigrants (such as sports or recreational associations) and the so-called "second generation". Only the first part of the quantitative analysis uses responses of such immigrant organizations (namely local immigrant associations' umbrella organizations in German cities, foreigner councils (Conseils des Résidents Étrangers in France,) and city-level branches of organisations like the Ligue Internationale Contre le Racisme et l'Antisémitisme (LICRA), Mouvement contre le racisme et pour l'amitié (MRAP), Association Solidarité avec Tous les Immigrés (ASTI), Comité inter mouvements auprés des évacués (La Cimade), Ligue des Droits de l'Homme, or SOS Racisme). Besides, key results presented in this article draw on survey answers on how other actors collaborate with immigrant associations and councils (see appendix). 3. For more information see (Moutselos et al 2017, p.10). 4. Team members are: Christian Jacobs, Christine Lang, Julia Martínez Ariño, Michalis Moutselos, Maria Schiller, Karen Schönwälder, Alexandre Tandé, Lisa Szepan. 5. For more information on the sample of cities see (Moutselos et al 2017, p.10). 6. We excluded individuals who act without an institutional base and less established initiatives. For a detailed description of the sampling procedure see Moutselos et al Technical report, p.10ff. 7. Of which 423 in Germany and 237 in France completed the network question of the questionnaire. Thus, in this article n = 660. Data of two cities were excluded from the multilevel analysis due to low response rates at the individual city level (therefore n = 590). 8. By migration-related fora we mean local policy fora in which issues related to immigration and cultural diversity are discussed. 9. All possible combinations off independent variables were tested, showing similar results than the full model. It's only worth mentioning that the inclusion of the variable concerning participation in local policy fora is what makes the country variable turn into a non-significant predictor. 10. This measure expresses the proportion of the variability in the outcome attributable to the units at the aggregated level (Field 2009) -the cities, in our case. Following Snijders and Bosker (1999, 224), the intraclass correlation coefficients (ICC) have been computed using the formula: r 1 = s 2 /(s 2 + 3.29), where 3.29 = p 2 /3. In other words, in this study it shows that overall there is almost no variation in establishing intense collaborations with immigrant advocacy bodies explained by differences between cities, but almost exclusively by differences within cities. 11. In Mannheim, there had been several incidents in the past where conflicts from the Middle East had spilled over to conflicts between immigrant groups in the city. City politicians and officials accorded high importance to such declarations as a means to prevent such conflicts from happening and to have tools to bring actors to the table and remind them of their responsibility for protecting a peaceful coexistence in the city. Table A1. List of the sampled cities and their characteristics regarding population size, percentage of foreign-born inhabitants, percentage of leftwing voters in the last local elections and percentage of surveyed actors participating in local policy fora.