Barcelona: municipalist policy entrepreneurship in a centralist refugee reception system

This article is a case study of Barcelona’s policy entrepreneurship in innovating refugee reception. In a context of a highly centralised and increasingly dysfunctional Spanish refugee reception system, it traces how Barcelona, after declaring itself a city of refuge, aimed to change refugee reception into a more decentralised and bottom-up system with a major role of cities. Following Minstrom and Norman, the article reconstructs the city’s policy entrepreneurship and shows how the city 1) led by example in developing innovative local policies; 2) developed political arguments for change from a municipalist stance in coalition with other cities at national and EU level; and 3) worked its arguments into more technical discussions on the future of integration funding in the EU. On the basis of these developments, the article reaches two main conclusions. First, while the different registers of policy entrepreneurship cannot be directly connected to any concrete change of the multi-level setup of refugee reception, they do represent a significant step forward in cities’ self-organisation beyond ideological lines and a factor contributing to a stronger recognition of cities in refugee reception in Europe. Second, while the literature on multilevel governance has an in-built bias towards more coordinated and harmonious relationships across tiers, the case of Barcelona shows that more attention should be paid to conflict and change.


Introduction
By late August 2015, at a time when the number of refugees arriving in Europe was growing exponentially along with the number of fatalities occurring during the attempt, Barcelona's mayor Ada Colau proposed creating a network of refuge cities. "We want", she said, "cities committed to human rights and life, cities that we can be proud of". In the same message Colau criticised the cynicism of "part of Europe" and made an appeal for change: "Europeans, open your eyes. There will never be enough walls or barbed wire to stop this. Or tear gas or rubber bullets. Either we deal with this human drama using the capacity to love that makes us human, or we will end up dehumanised. And there will be more deaths, many more. This is not a battle to protect us from others. Right now there is a war against life" (Ajuntament de . In this article, we want to trace the efforts of the city of Barcelona to become a "city of refuge" that followed this statement of its mayor as an example of municipalist policy entrepreneurship in refugee reception. While in the area of immigrant integration, both Autonomous Communities and cities like Barcelona could play their part in policy development (see Garcés-Mascareñas 2014), in asylum, the centralised Spanish reception system did not foresee any such role for them. When asylum became an issue in Spain in the summer of 2015, this immediately turned into a conflict between the lower tiers of government claiming to share responsibilities and the national government reluctant to do so.
This struggle for competences was partly related to political ideology, but also went beyond a simple left-right division: during the period from 2015 to 2019, which we analyse in this article, Spain, Catalonia and Barcelona were governed by three ideologically antagonistic forces: a conservative government of the Partido Popular at national level, which was replaced in 2018 by a Socialist government; a coalition of the independentist parties ERC (centre-left) and PDeCAT (centre-right) governing Catalonia; and the platform of post-15-M citizens' movements Barcelona en Comú governing the city of Barcelona. The constitutional conflict about Catalan independence certainly added to the potential for ideological conflict. In addition to that, at the local level, the proximity of Barcelona en Comú to social movements (see, for instance Eizaguirre, Pradel-Miquel, & García, 2017), human rights and pro-refugee groups led the Mayor of Barcelona to channel the tangible pro-refugee sentiment in local civil society to the national government, criticising its passivity. While this activism was inspired by the city's municipalist outlook to radically challenge the status quo from the local level, in the course of the conflict, Barcelona also mobilised city-specific, non-ideological arguments, a sort of policy-coordination common-sense, and built heavily on alliances with other cities at national and international level (see Broadhead in this special issue), which were more founded on city-ness than on political colour.
In the context of this special issue, this article aims to understand not only how cities innovate in the policies falling under their responsibility, but also to what extent they have the capacity to change the conditions for innovation as defined by a governance setup perceived as unfavourable to pursue their political priorities. Our goal is to examine how the city council articulated its political claims in the field of asylum reception, with a particular attention to the ideological, legal and technical arguments used; which interlocutors and allies it involved; and which outcomes were reached in terms of challenging and redefining the status quo in asylum reception policies.

Methodological approach
To respond to these questions we build on research done in the context of three different projects. In 2016 research was done to compile data on local reception policies in Barcelona for the IMISCOE research group on the role of municipalities in giving response to the 2015 migratory crisis. Soon it became evident that, beyond local reception policies, the case of Barcelona was of particular interest for its strategy at the international level, using international city networks, bilateral city-to-city relations and actively participating in international events. This is what attracted our attention in 2016 and 2017. In 2018 we conducted new research, this time under the umbrella of the H2020 project CEASEVAL and with the specific aim to understand the multilevel governance of reception policies from a comparative perspective. Our research focus did thus somehow change over time, and so did the context in terms of a rising number of asylum applications and new policy approaches. Instead of being a limitation, this allowed us to have a diachronic and multidimensional perspective.
Our analysis is based on two types of primary data. First, we analysed policy documents, reports by the main stakeholders, official data and newspaper articles from 2015 to 2018. Second, we conducted 25 interviews with politicians and civil servants at different administrative levels and representatives of social and civil society organisations, mostly in Barcelona. In both cases, we paid particular attention to how the problem was framed by the different actors at different levels and in which way ideological, constitutional and technical arguments shaped actors' behaviour and interactions.
The structure of the article is as follows. The first section provides a framework for our analysis based on the literature on multilevel governance and policy entrepreneurship. We then provide a reconstruction of Barcelona's municipal entrepreneurship along three main lines, which partly reflect Mintrom and Norman's categories of policy entrepreneurship: the first is about the city leading by example and developing local responses in asylum in the context of a top-down and overburdened Spanish system for refugee reception. The second is about developing arguments for changing the governance framework and communicating them together with like-minded cities in Spain and Europe. The third is about the city backing up its political argument by its participation in technical discussions on the future of integration funding in the EU Urban Agenda Partnership. We will thus show that while the city's policy entrepreneurship has a strong political component, informed by the municipalist philosophy of its leadership, it is also nourished and complemented by technical exchanges that the city has been engaged in for a long time through other agents.

Multi-level governance and city policy entrepreneurship
The relationship of cities with other tiers of government has received increasing attention in the framework of multi-level governance (MLG). According to Caponio and Jones-Correa (2018), MLG encompasses (1) the involvement of different levels of government (vertical dimension); (2) the involvement of non-governmental actors (horizontal dimension); and, the most contested third elements of "the emergence of complex, heterogeneous, and non-hierarchical networks" of actors. Regarding the vertical dimension, cities become particularly relevant whenever they go beyond the role of simple passive implementers and actively interact with national or regional policies. In the field of integration policies, cities tend to hold competencies in integration policies or in areas that are more or less directly relevant, such as social services, housing, employment etc. As for the horizontal dimension, cities represent a particular interest for MLG due to the stronger presence of civil society organisations in local governance, certainly in integration policies. According to Caponio and Jones-Correa (2018), the nature of the relationships between different tiers of government (3) is probably the most contested element of MLG. In particular the parallel use of MLG from a normative perspective, in which policy coordination across tiers is seen as element of "good governance", and from a descriptive perspective to analyse a particular governance setup involving several tiers, complicates communication.
Continuing with the second, analytical perspective on MLG, Scholten's (2013) ideal-types of modes of governance in multi-level settings provide a simple scheme to characterise governance setups, which we will draw on in the following. Scholten distinguishes (1) topdown (in his words, "centralist"), (2) bottom-up ("localist"), (3) de-hierarchised, ("multilevel") and (4) "decoupled" relations, which are presented as typical outcomes of different constellations of interests and policy agendas across tiers of government. What is maybe not appropriately reflected in most theories and categorisations of MLG is that, at least in immigration governance, conflicts and a de facto decoupling between local and national policies are frequent (Zapata-Barrero, Caponio, & Scholten, 2017) and represent probably more of a normality than cases corresponding to the de-hierarchised and well-coordinated MLG-ideal type. This is where the both normative and analytical approach to MLG is problematic, as it might suggest somehow a trend towards more coordination, while in reality, from the enforcement of US-immigration legislation to the redistribution of refugees in the EU, quite substantial conflicts between tiers of government exist. In addition to ideological differences between tiers being governed by different parties, it is plausible to assume that what Spencer (2018) calls "tier specific" ways of perceiving and framing immigration between cities and the state play a major role in explaining such conflicts and tendencies of decoupling. This includes different ways in which problems manifest themselves (e.g. in the form of illegal border crossing or of destitute people sleeping rough) and, even more fundamentally, different conceptions of the demos of the city versus that of the state (Gebhardt, 2016).
Moving on to the specific role of cities in such conflicts over MLG, parts of the literature on the "local turn" in immigration (Caponio & Borkert, 2010) have indeed claimed that cities have become "increasingly entrepreneurial" by "pushing their own policy ideas onto the national agenda" (Scholten, 2013), "taking policy directions very different from their national governments" (Scholten & Penninx, 2016), becoming more proactive in engaging at international level (Caponio, 2018) and conducting venue shopping (Giraudon, 2000) at supranational policy venues so as to circumvent national governments. At the same time, evidence on how exactly cities become entrepreneurial in their interaction with other tiers of government and try to change governance setups into a more coordinated direction is still scarce (for one recent exception, see Scholten, Engbersen, Ostaijen, & Snel, 2018). A tendency to generalise on the role of cities in integration governance has meant that cities have often been idealised (e.g. as more inclusive) or banalised as mere parts of the state, instead of looking at what exactly they were doing.
In order to better understand policy entrepreneurship of cities, it is helpful to take one step back and look at the work on this concept beyond the city/immigration context. According to the overview by Mintrom and Norman (2009), policy entrepreneurs are "advocates of policy change" who operate closely to innovation, as "policy entrepreneurship is most likely to be observed in cases where change involves disruption to established ways of doing things" (Mintrom & Norman, 2009). Although most of the scholarly work on policy entrepreneurship has looked at individual entrepreneurs belonging to the political and administrative level of government or to particular interest groups, the concept has also been extended to institutions (e.g. Moravcsik, 1999 for the European Commission or Perkmann, 2007 for Euroregions).
From their overview of the literature, Mintrom and Norman (2009, pp. 651-654) identify four central elements of (successful) policy entrepreneurship. The first element is social acuity, which is about the presence, relationships and connectivity of policy entrepreneurs within policy networks and their understanding of other actors in their policy context. This allows seizing windows of opportunity to drive forward policy change. The second element is providing a convincing definition of the problem that should be tackled and that provides arguments for the policy change, e.g. failures of current policies. The third element is building teams and coalitions that put advocacy for change on a broader basis. The fourth element is leading by example, for instance by developing pilot projects demonstrating how policy change could work. These four categories of policy entrepreneurship can be specified for our case in the following way: identifying a governance problem in refugee reception and providing arguments for why it has to be addressed; demonstrating commitment to act upon the problem by developing innovative policies; entertaining relationships with a wide array of relevant actors to communicate the problem; and building alliances with like-minded actors to advocate for change. A key question in this context is what the main motivations behind and returns of city alliances are and to what extent local initiatives at the international level contribute to reconfigure city-state relations. In a recent article, Caponio (2018) claims that city networks play, first and foremost, a symbolic function by contributing to identitybuilding and political legitimation, but also to a city's positioning.
Leading over to our case study, the policy entrepreneurship of Barcelona to change the governance setup in refugee reception in Spain was strongly influenced by a wider "municipalist" philosophy of the city. This idea, which goes beyond a localist or decentralisation agenda, and instead sees the city "as a 'strategic front' for developing a transformative politics of scale" (Russell, 2019), has been a key concept for the political leadership of Barcelona since 2015, in areas including, but also going beyond migration and asylum policies. Municipalism is therefore the concrete manifestation of policy entrepreneurship in MLG in our case study.
Leading by example: a local asylum reception system The Spanish asylum system is highly centralised in the sense of the classification developed by Scholten (2013). It is a state competence to process asylum applications and to accommodate asylum seekers mainly via social organisations working under its instructions. The system foresees no coordination between the state and the regional or local level, which explains why, up to 2015, municipalities had limited their actions in asylum to offering legal advice to potential asylum seekers, directing them to state reception programmes and at best including them, like Barcelona, as target group of local services for newcomer orientation. Given the low numbers of asylum seekers that had come to Spain until 2015, this centralised governance had not posed a major political problem, compared to, for instance, conflicts over the management of irregular migration or migrant integration.
All changed in 2015, with the expected increase in the number of refugees arriving in Spain through the relocation quota from Italy and Greece. It was to this expected growth that the city of Barcelona reacted. The first programmes were urgently prepared in September and October 2015 as the feeling at the time was that the first refugees resettled from Italy and Greece would be arriving immediately. The City Council developed the "Barcelona City of Refuge" Plan, with the objective of preparing the city to receive, assist and provide the necessary services for refugees and guarantee their rights'. Coordinated by the City Council's Technical Director (who is part of the Mayor's Office) and supported by a broad volunteer network managed by social organisations, the programme followed the model of local emergency plans in other areas.
But the asylum seekers meant to be relocated from Greece and Italy did not arrive as planned: 2 years after the relocation agreement, only 8% had been relocated to Spain. In parallel, the number of asylum applications in Spain increased from 14,881 in 2015 to 31.120 in 2017, 54,065 in 2018 and 118,264 in 2019 (Comisión Española de Ayuda al Refugiado, 2020). These asylum seekers came initially predominantly from Ukraine, Venezuela and Algeria and later from Venezuela, Colombia and Central American countries. Despite asylum seeker reception being the responsibility of the national government, the increase in the number of applications and the consequent saturation of the state reception system led Barcelona City Council to transform its initial emergency programme for those to come to a more structural reception system for those already in the city.
This meant, on the one hand, increasing the number of temporary accommodation places in council guesthouses and schemes. On the other, municipal authorities designed a parallel reception plan to complement the state reception system. In particular, in order to accommodate those asylum seekers who did not enter yet, fell out of, or already left the state programme, the City Council set up the Nausica programme, which was granted over a million Euros between 2016 and 2017. The programme provided two instruments: one was a complex of social interventions, which included economic benefits to address situations of economic vulnerability, a programme for increasing employability and measures to respond to specific social and psychological needs; the other instrument additionally included temporary housing places.
The refugee reception policies led by the city of Barcelona were financed by the city's own resources. In total, in the first 3 years after its adoption, the city spent about 6 m Euros for its refugee strategy (Interview Politician Barcelona, March 2018). Although it was meant to address the gaps of the state refugee reception system, it did not actively integrate with it. In addition to lack of funding, the perception in the City Council was that the state had made no efforts to support and coordinate with the city's policies. A policy maker working for the city stated: "since 2015 we have had just two meetings with the state. They receive us very politely but nothing more […]." The policy maker acknowledged though that the change of government in summer 2018 (from the right wing Partido Popular to the Socialist Party PSOE) brought a different attitude from the side of the state: "With the People's Party we have been living in a desert, it's been terrible. With the PSOE governments it was much better. The Popular Party does not believe in local governments, it has nothing to do with the political situation here in Catalonia (Interview City Council Manager Barcelona, December 2018).

The politics of city alliances
When Barcelona City Council launched its refugee policy plan, or, more precisely, as part of it, the city also started to build networks with similar-minded cities. The first initiative was the Spanish cities of refuge, ciudades refugio. Following Mayor Ada Colau's declaration in September 2015, Barcelona was soon joined by other Spanish cities proclaiming themselves cities of refuge, committing to host refugees and making claims visa-vis the Spanish government. In parallel to this, Barcelona also engaged with other cities in Europe. This happened in different constellations, but mainly through existing partnerships within Eurocities, and the initiative Solidarity Cities, which was co-founded by Barcelona in October 2016 based on prior exchanges between Athens, Barcelona, Berlin and Amsterdam. As we will show in the following, both networking activities served to develop common political messages, which in particularly targeted the Spanish government and the European Union.

Spanish cities of refuge
In early September 2015, while the Spanish government still opposed the relocation quota proposed by the European Commission, Colau declared Barcelona an "open city" or "city of refuge". A week later, 55 Spanish city councils had declared themselves cities of refuge, too. Joining Barcelona were the two other biggest Spanish cities, Madrid and Valencia, as well as others such as Cádiz, Zaragoza, A Coruña, Córdoba, Huelva, Palma de Mallorca, Toledo, Valladolid, Vitoria and Pamplona. In concert with these other cities, Barcelona engaged in an exchange of public declarations with Mariano Rajoy, then President of the Spanish Government, which went on for 2 years (from September 2015 to September 2017). In this period, the cities created and used different events to engage with the government. Such events were the Vatican summit "refugees are our brothers and sisters" in December 2016, the massive pro-refugee demonstration "Volem acollir" (We want to welcome) in Barcelona on February 2017, a formal declaration made by 25 Spanish cities in June 2017 under the title "Cities and the reality of refugee reception in the Spanish state" (Ajuntament de Barcelona, 2017)  The cities demanded, firstly, to monitor the government's (lagging) implementation of the EU relocation plan. They also asked the central government to respond and comply with its obligations. In one case, Barcelona proposed a pilot for city-led relocation with Athens to help the state to meet the relocation goal. This is where the cities of refuge initiative touched directly with its European counterpart Solidarity Cities (see below). Secondly, the cities claimed resources from the state to meet their political commitment in refugee reception. In particular, they referred to the fact that none of the 521.7 million euros allocated to Spain by the European Union through the Asylum and Migration Fund had reached the cities. Already the first letter to Rajoy from September 2015, signed by Ada Colau, said "The city councils need funds and we know there are resources", asking for more transparency in the allocation of these funds (Colau Ballano, 2015). Thirdly, the cities demanded better policy coordination with the state, therefore urging the government to meet them. For instance, the letter signed by Barcelona, Madrid, Valencia, Zaragoza and A Coruña after the Vatican summit asked the government for a meeting at "highest level" in the "shortest delay possible" to "unblock" the problems of a lack of coordination in refugee reception. Colau's communication after the "Volem Acollir" demonstration in February 2017, which gathered around 500,000 people, made the point that "nobody of this government ever wanted to meet the city of Barcelona, while the European Commission, the United Nations and the Vatican did".
The arguments that cities used to back their claims appealed to their status as democratically elected governments representing a significant number of citizens. For instance, after the "Volem Acollir" demonstration, Colau said that Rajoy was "not only ignoring the thousands who went out to the streets", but also turning his back on the mayors of five of the biggest cities of Spain, representing 6.5 million people. From a more legal point of view, the cities built their claims on a human rights obligation to provide asylum, and on their own competency of providing support for vulnerable people (Carmena, Colau, Ribó, Santisteve, & Ferreiro, 2016). They also developed an efficiency argument by claiming that the lack of coordination would force cities to spend a "great amount of resources" for the accommodation of asylum seekers outside the state system, which could be avoided with a "minimum of coordination" based on "institutional loyalty" between central and local administration (Ajuntament de Barcelona, 2017).
The arguments of the state government rebutting the cities' claims took first and foremost a legalistic form, stressing that asylum was an exclusive competence of the state. Rajoy's response to the first letter in September 2015 already accused cities of exceeding their competences by developing their own initiatives. When Colau asked for city-led relocation of refugees, the president further sharpened his tone reminding the government's "exclusive competence" in the management of resettlement and relocation. In the statement of the state secretary after the Vatican Summit letter this argument was further developed: "The state, and not other public administrations, ensures that these international obligations (regarding asylum) are met and that the support is equal across the national territory" (Bermúdez de Castro Mur, 2017, translation by the authors). That way, the local initiatives were almost depicted as a threat to territorial cohesion. Related to this, there was an efficiency argument made by the state, according to which cities were invited to support the state's efforts instead of developing their own initiatives. The cities' policies were characterised by Rajoy as "patchy" and "short term solutions", and contrasted with the state's systematic efforts, in particular in curbing migration flows. In all its communications, the government reassured that it had the situation under control, that it would "fulfil its pledge within the EU relocation scheme" (e.g. Rajoy Brey, 2016) and that it was taking all necessary efforts to coordinate with lower tiers of government (Bermúdez de Castro Mur, 2017).
While initially the letters sent by Barcelona to the other cities of refuge received an immediate response by the President of the Spanish Government, delays increased as the dialogue went on and the reply came by the state secretary for public administration instead of the president (Roberto Bermúdez de Castro Mur, 2017). Eventually, the government stopped answering altogether and refused to receive cities at the national parliament: when in September 2017 a delegation of cities of refuge went to the parliament to mark the deadline foreseen in the EU relocation plan, it was only received by representatives of the left-wing opposition party Podemos.
What clashed in this exchange of arguments between cities and the state government were different ideas of legal duties (cities' insisting on human rights obligations and the government on its formal competences regarding asylum) and different ideas of the need for policy coordination (which, according to cities had to include them, whereas the government, again, insisted that its top-down or decoupled approach was the most efficient). As a result of these opposing views, the concrete outcomes of the Spanish network of cities of refuge in terms of changing policies were limited. At the same time, the crisis of the Spanish refugee reception system increased, with a backlog of pending asylum claims building up since 2015 to reach over 40,000 in 2018 (Comisión Española de Ayuda al Refugiado, 2020). In spite of this, the Spanish city network can be seen as a success in organisational terms, as it managed to extend its coalition to include city councils of all political colours. This was most obvious the case in the 2017 "Barcelona Declaration" with the illustrative title "Cities and the reality of refugee reception in the Spanish state", which raised "city-ness" over political ideology and voiced an alternative position to the government's.

EU-wide city-networking
In the same month of September 2015 in which the Spanish cities of refuge network was launched, Barcelona also participated in a first city-initiative at EU-level in which Ada Colau and the mayors of Paris, Lesbos and Lampedusa published a manifesto asking for more state support: "We are ready to become places of welcome ... We have space, services and, most importantly, the citizen's backing to do it. Our municipal services are already working on reception plans to ensure bread, shelter and dignity to all those fleeing war and hunger. The only thing that is missing is support from the states" (" Barcelona, Paris, Lesbos and Lampedusa," 2015). The letter was also endorsed by Madrid, Valencia, Zaragoza, A Coruña, Cádiz and Santiago de Compostela, six self-proclaimed "cities of change" forming the hard core of the Spanish cities of refuge initiative.
From the outset, EU-level city-networking had the double aim of building alliances and exchanging experiences, which is illustrated by Ada Colau's visits to Leipzig and Munich in 2016 (" Barcelona busca referents," 2015). Networking also included building contacts with refugee arrival cities at the borders of the EU and beyond, to provide technical and logistical support from city to city. In March 2016, a new type of cooperation was envisaged with an agreement for a pilot city-to-city relocation scheme of 100 refugees from Athens to Barcelona. While this initiative was denied by Mariano Rajoy, it marked the beginning of a new level of institutionalisation of city refugee policies within the EU under the title 'Solidarity Cities'.
Within Eurocities, the network of major European cities, 'Solidarity Cities' was initiated by Athens and included Barcelona, Berlin and Amsterdam as founding members. The initiative presented itself as a city-to-city mechanism of solidarity, consisting of mutual help and city pledges for the relocation of refugees. According to the coordinator of the Barcelona Refuge City plan, the message of Solidarity Cities was "if cities can work together, states can, too!" (Interview Politician Barcelona, February 2016). So Solidarity Cities aimed to mirror the failing relocation agreement between EU member states through a functioning trans-local cooperation based on solidarity. The presence of Solidarity Cities like Barcelona, Warsaw and Gdansk in reluctant asylum states like Spain or Poland also aimed to demonstrate that this reluctance was not shared across the whole territory.
At the same time, Solidarity Cities provided an attractive platform to engage with EU institutions. On 5 April 2016, at a meeting in Brussels organised by Eurocities, the mayors of Barcelona, Athens, Amsterdam, Ghent, Helsinki, Berlin, Leipzig, Malmö, Paris and Rome met with the Commissioners for Regional Policy, Creţu and for Home Affairs, Avramopoulos. At this gathering, which was widely covered by the media, Ada Colau had the opportunity to again denounce the Spanish state's "immoral" management of the refugee crisis and lacking coordination with cities. In her blog, Colau stated that true multilevel governance needed financing. After again pointing out that the Spanish government refused to explain how it managed European Union funds for refugees and asylum, a situation she described as "anomalous" and "dysfunctional", Colau asked for a "percentage (of European financial support) as direct funding to the municipalities". In response to cities' claims, the Commissioner for Migration, Home Affairs and Citizenship, Avramopoulos, recalled that the interlocutors of the European Commission were the governments of member states and that European aid to local authorities "is always through the member states" ("Diez alcaldes europeos piden," 2016). However, he also asked the member states to "listen to the local governments" because "internal cooperation and coordinated efforts could give impressive results".
While there were no immediate outcomes of the political meetings between Solidarity Cities and EU Institutions, just like in the Spanish network, its political success in terms of making visible and promoting a city-position beyond the status quo should not be neglected. This is illustrated by the presence of the initiative in mainstream media, e.g. the Guardian publishing on world refugee day, 20 June 2016, an open letter of Eurocities' executive committee, which includes Ada Colau, to the presidents of the Council of the European Union, the European Parliament and the European Commission: "Now is the time to put our shared European values of solidarity, humanity and dignity to the test. Cities are where the integration of newcomers will succeed or fail. We, as leaders of major European cities, want European leaders -President Juncker, President Schulz, President Tusk and national leadersto work with us, not only by acknowledging our challenges but also with concrete actions such as direct financial support to cities" ("Refugee challenge for cities," 2016).
Solidarity Cities allowed embracing new opportunities to be present in the public discourse on migration and to spread a city-narrative that was distinct from the mainstream discussions between states. This is particularly the case with the topic of city-led relocation that Barcelona promoted within Solidarity Cities, which has gained importance in European discussions more recently, although it is difficult to say in how far this is a direct consequence of the initiative: together with initiatives on city-led resettlement such as the SHARE network, 1 cases of city-relocation pledges that were agreed on by national governments, such as that of Lisbon, and a growing number of civil society initiatives for communitysponsored resettlement and local safe harbours, Solidarity Cities helped spreading the idea of city-led refuge. This culminated in the French president Macron suggesting in a speech at the European Parliament in April 2018: "Je propose ainsi de créer un programme européen qui soutienne directement financièrement les collectivités locales qui accueillent et intègrent des réfugiés" 2 (European Parliament, 2018). While there is no formal recognition yet of city-led admission, relocation or resettlement, the discussion has significantly moved on in this area.

The technical route of the urban agenda partnership
In addition to shaping a political message on the role of cities in refugee reception through Solidarity Cities, Barcelona also continued its work through policy venues of a more technical nature, building on the more than a decade-long work of the city through Eurocities. While the more technical work involved different participants, they were neatly connected and coordinated with the political activities (Interview City Council Manager Barcelona, 2018). Technical knowledge sharing and political commitment legitimised each other.
The Urban Agenda Partnership on the inclusion of immigrants and refugees exemplifies well this technical side of city entrepreneurship. In early 2016, the Dutch Presidency of the EU and the European Commission's Directorate General (DG) for Regional Policies, the most city-friendly directorate of the European Commission, set up Urban Agenda Partnerships in 13 areas reaching from air quality to public procurement. Each partnership was formed by a limited number of city and member state representatives together with relevant EC DGs and think tanks to improve the EU policies for cities from the perspective of funding, regulation and knowledge sharing. With their multi-level governance setup, the partnerships can be seen as an experimental tool to improve policy coordination with cities.
The partnership for the inclusion of migrants and refugees included the cities of Amsterdam, Athens, Barcelona, Berlin and Helsinki -cities represented in Eurocities' migration working group and in Solidarity Cities, in addition to the member states Denmark, Greece, Italy and Portugal, and the DGs for Regional Policies, Home Affairs and Employment. From the outset, Barcelona's interest in the Partnership was the redefinition of the financial channels within the European Union and laying the foundations for true multilevel governance (Interview City Council Manager Barcelona, 2016). Initially, the clash between the EU policy world and the city's reality was perceived as immense: a policy maker from the city council who participated in the partnership on behalf of Barcelona explained that at one partnership meeting, when urged by a EC representative to contact the national AMIF managing authorities, he answered: "Excuse me. My managing authority hasn't spoken to us for the last five years ( …). They have never, never, never been in touch with us." He further commented that "The guy [from the EC] didn't get it. They're very rigid. They live in a world where they think that everything works because there's a decree. But the states do what they damn well like. There are some countries in which the relationship between levels works, where institutional loyalty exists. Not here. In Spain, it's a disgrace!" In spite of the city representative's initial scepticism towards the partnership, it did produce a tangible outcome in line with the city's objectives: in March 2018, the Partnership published, under the lead of Eurocities, its recommendation paper for improving cities' use of and access to integration funding as input for designing EUintegration funding post 2020 (Urban Agenda of the EU, 2018). The problem analysis of this document included "little or no involvement of municipalities" in integration funding; and the "non-recognition of cities as target group of funds and as partners by managing authorities". It thereby reiterated, in greater detail, some of the messages developed in Solidarity Cities and by the Spanish cities of refuge and recommended experimenting with direct access of cities to funding.
Building on these recommendations, the 2018 draft EC regulation for the Asylum and Migration Fund after 2020 took up on the narrative of the recommendation paper, agreed on the need for a stronger outreach to cities and foresaw to reduce the share of the fund that is implemented by member states from 88 to 60%, leaving the EC in charge of a "thematic facility" worth 40% of the total budget, in which cities would count among the explicit target groups (European Commission, 2018, Article 9.6).
The discussion on the next generation of integration funding is still underway, which does not allow drawing conclusions on whether the presence of Barcelona and other cities in the EU has led to tangible results. But the fact that the "city imprint" in the draft EU regulation for the European Migration and Asylum Fund is stronger than in its predecessors, in spite of the arguably stronger grip of national governments on the topic of integration, can be seen as an indication for a wider shift in favour of cities.

Conclusions
This article analysed the city of Barcelona's policy entrepreneurship towards an increased role and stronger support for local governments in the highly centralised governance of refugee reception in Spain. In the context of multi-level governance, we characterised this case as one of many where systematic conflicts over shared governance between cities and state governance arise. This conflict was particularly extreme as it was fought out between two ideologically opposed tiers of government with opposing views on asylum: a national government prioritising curbing refugee flows that was highly reluctant to share responsibilities; and a city government with a tradition of promoting equal rights and opportunities and recognising cultural diversity in the Spanish context and a municipalist agenda of aspiring a wider change towards a more welcoming and human rights based reception system. We would argue that such conflicts do not receive enough attention in studies on multi-level governance, and suspect that this might be due to an in-built bias of the concept towards ever more coordinated and horizontal relationships. In recognising the conflictive reality and shifting the focus from the level of government holding the formal competence to who de facto participates in the design, implementation but also political discussions around a particular policy, the picture becomes much more complex. In our case of the Spanish asylum system, what may appear as a top-down, centralist governance model shows important elements of de-coupling and is, at the same time, challenged from more localist, bottom-up positions. We believe that such complex conflicts and how they contribute (or not) to a change of the governance set up deserve more attention.
The many registers that Barcelona pulled from its municipalist stance to change the governance setup in refugee reception and achieving a more welcoming reception system fit quite neatly with the four elements of policy entrepreneurship of Minstrom and Norman: the city provided a definition of the (governance) problem, stating that the centralised Spanish refugee reception system was dysfunctional, inefficient and undemocratic (in leaving out the cities), that it did not respect human rights and did not comply with international obligations such as the EU relocation plan.
In the absence of policy coordination and support from the side of the central government, the city led by example by developing pilot projects, such as a local reception system to accommodate those left out by the state reception system by providing temporary accommodation and measures for fostering employability and welfare of asylum seekers. These measures were developed without any coordination with or financing by the Spanish government and correspond to what Scholten (2013) defines as decoupled governance arrangements.
Minstrom and Norman's element of building teams and coalitions that put advocacy for change on a broader basis was one of the key activities of Barcelona city council's policy entrepreneurship. The city developed a twofold strategy: leading the Spanish informal network of cities of refuge, to pressure the Spanish government into more coordination, and participating in the European Solidarity cities initiative, which targeted both national governments and EU institutions. Across political divides both networks managed to forge a common position claiming above all an acknowledgement of the role of cities in refugee reception as well as better and direct financial support. These political activities were accompanied by more technical relationships for reforming integration funding through the EU's Urban Agenda Partnership.
In terms of social acuity, we could see how the city's capacity to meet representatives from other (non-city) governments was more developed in the activities at European level than in Spain, where such contacts were prevented by strong ideological conflicts and a lack of inter-institutional loyalty and cooperation. The city's connectedness in Europe, which allowed it seizing windows of opportunity, is a fruit of its decade-long work in European contexts, and of it working through Eurocities, which provided a rather institutionalised channel for such engagement. This led, for instance, to opportunities to meet European Commissioners and participate in the development of recommendations for reforming integration funding. The European engagement goes beyond a mere "venue shopping" logic: more than eluding an unfavourable national context by "going to Europe", the city council combined different venues from the outset, and its European engagement did not just represent a complementary path.
In drawing so heavily on coalitions with other cities, Barcelona's municipalist policy entrepreneurship has been critical in developing a common "cities" position on refugee reception that was self-confidently positioned against the status quo in Spain and Europe. Barcelona could draw on the support of other cities to increase the weight of its arguments, and of its political agenda. In this regard, our findings nuance those of Caponio (2018) for different cases of city networking in that Barcelona's strategy went far beyond the purely symbolic dimension of identity-building, political legitimation and international recognition of cities, and was successful in putting its demands onto the political agenda, in particular that on the future of EU integration funding. At the same time, what in the end triggered the most radical change in asylum governance in Spain were events beyond the city's reach: in 2018 a judicial appeal of the Catalan government against the Spanish government from 2016 which alleged that asylum reception was a competence of the autonomous communities was ruled in favour of the former, and finally opened the door to restructuring the Spanish refugee reception system towards a more decentralised and coordinated approach. The role of cities in this coming reform still remains open.