Re-compartmentalizing local policies? The translation and mediation of European structural funds in Sweden

ABSTRACT The local governments of EU member countries are attracted to the possibility of receiving EU funding. However, as the governance structures of EU funds are complex and dynamic, municipalities are increasingly drawing on the knowledge and resources of ‘EU experts’ who mediate and provide project support. This article contributes to our understanding of how EU cohesion policy is translated through EU projects, with a specific focus on the processes of preparing and applying for project funding. Drawing on education policy, this study analyses a tool which has been developed to facilitate and increase the number of EU projects in the Swedish region of Scania. The analysis shows that regional mediation – and the ambition to reframe local policies into EU projects – entails substantive as well as organizational changes in two aspects. First, the policy content shifts from the realm of education policy to the realms of collaborative development policy, social cohesion and innovation, and second, the translation entails an organizational shift from permanent public education administration to temporary project organizations. These processes are conceptualized as the re-compartmentalization of local policies.


Introduction
How can local government organizations navigate in the context of EU project funding? And how is knowledge mediated and translated from the EU to municipalities in these processes? Project funding and project organizing have become an essential and integral part of EU policy-making on different levels during the recent decades (cf. Büttner and Leopold 2016, 42). A large percentage of the EU's budget is managed through various types of project funding systems, and a large proportion of these funds are implemented at the local level (cf. Büttner and Leopold 2016;Godenhjelm, Lundin, and Sjöblom 2015). Nevertheless, research on Europeanization tends to focus on member states adapting to and implementing various types of EU output on the national level (Guderjan 2012). On the local level, the literature is scant (cf. Sanchez Salgado 2013;Van Bever, Reynaert, and Steyvers 2011), in particular, regarding specific processes of local mediation and the translation of EU cohesion policy through EU funding.
EU Cohesion Policy is one of the EU's main investment tools, and it is delivered primarily via funds, described by the European Commission on their homepage as 'the policy behind the hundreds of thousands of projects all over Europe' (homepage as of May 25 2018). The structural funds have been described as one of the main domains of multilevel governance (Bache and Chapman 2009). Even though each fund is different in content and scope, there is a logic and an established implementation structure of project financing that transgresses the different policy fields (Büttner and Leopold 2016, 49). In practice, this has come to entail processes of not only changing the power distribution between levels of policy-making, but also the gradual development of new 'roles for both existing and newly-created institutions and bodies' (Stephenson 2013, 828). In response to the governance architecture of the EU, local government organizations have participated in the development of numerous new organizations, new professions, and policy instruments with the purpose of increasing EUfunding opportunities as well as facilitating collaboration, coordination, and network building (cf. Büttner and Leopold 2016, 42;Sanchez Salgado 2013). In this article, we study these governing processesand the development and prevalence of what could be described as EU expertsthrough an analysis of a method called EU project analyses (EPA). The EPA method or device is used by consultants and/or regional civil servants to map how municipalities can use EU funding in local development planning (Scania Association of Local Authorities (SALA) webpage, 2016). In practice, the EPA analysis entails a process whereby municipal documents such as budgets and local plans are surveyed and 'scanned' in an ambition to 'match' local priorities with EU project funding opportunities.
The overall aim of the article is to enhance our understandings of how European social cohesion policy is mediated and translated by regional EU experts through local processes of preparing and applying for projects from European structural and investment funds. Following this, we also investigate which (a) substantive and (b) organizational changes can be traced in these processes of translation.
The article uses examples from Swedish municipalities to argue that the increasing incidence of EU funding entails not only organizational but also substantive policy changes in local government. These multilevel processes give rise to the creation of new types of regional actors, here described as a form of EU experts, who are both enacted byand acting throughprocesses of mediation and translation. In terms of substantive change, the translation from the EU to Swedish municipalities requires the reframing of one policy area to another. We illustrate how education is reframed to include concepts such as 'skills development' and 'employability' as well as 'social exclusion' to 'match' (or suit) the conceptual framework of the EU Cohesion Policy. Here, we also relate our findings to previous discussions on the Europeanization of education policy (see e.g. Alexiadou and Lange 2015). In terms of organizational change, we argue that the process of 'matching' municipal goals into an EU framework entails processes of translations that, in turn, constitute an inherent shift in organizational form from permanent (mandatory) activities to temporary (project) organizations. These dual processes of change are captured in the concept of recompartmentalization, which is defined as a process whereby issues are not only reframed in processes of translation but also moved from one part of the municipal organization to another and thus dealt with through temporary projects instead of ordinary (permanent) operations. These organizational aspects will be elaborated below in the description of the Swedish municipal committee and sector division.
The article is structured in the following way. The first section outlines the main theoretical and methodological concepts. Here, the concepts of translation and Europeanization are used to describe the ambition to 'match' local policy goals into an EU framework to ensure funding, and the concept of mediation is elaborated to describe the regional organizations that conduct the EPA analysis. Thereafter, the empirical material is presented along with a description of the method of textual analysis employed. However, before the empirical analysis is presented, the empirical municipal context is described, including the Swedish municipal organizational setting and the concept of re-compartmentalization. The article concludes with an outline of the main results.

Europeanization as translation and mediation
The EU Structural Fund reform (1988) acts as the foundation for the EU Cohesion Policy (Dabrowski and John -Bafoil 2014, 355). The implementation and governing of the funds has been described as a dynamic process where the commission continuously introduce new governance modes (Tömmel 2016, 117), and an emerging governing context where municipalities are encouraged to align their activities to that of the EU (Bruno, Jacquot, and Mandin 2006;Godenhjelm 2016;Fred 2015).
Processes of local change in relation to the structural funds and cohesion policies can and have been analyzed in terms of Europeanization (see e.g. Verschraegen, Vanhercke, and Verpoorten 2011;Sanchez Salgado 2013). Theories that fall under the conceptual umbrella of Europeanization are generally employed to understand the forms and shapes of the EU's influence on its member states, including changes or adjustments in national institutions and policy-making as a result of EU governance (Cort 2010;Tonra 2015;Börzel and Risse 2003). Europeanization has been understood as a process of including all three elements up-loading, down-loading, and cross-loading. In this context, cross-loading refers to the learning between nations, up-loading refers to the processes of conveying national preferences to the EU level, and down-loading refers to the processes where EU-level directives, agreements or regulations are embedded on the domestic level, entailing a shift in content or scope (cf. Tonra). As Pia Cort (2010, 332) notes, these endeavors raise questions regarding both how 'processes translate into policies at national level and how they should be researched'. A body of literature has also discussed processes of Europeanization in relation to governing, where the focus has been on the relations between European integration and domestic policy-making (Alexiadou and Lange 2015). In what Radelli (2004, 5) describes as the so-called second generation of Europeanization studies, Europeanization is no longer seen as a specific form of theorization, but rather the concept can be understood as an '"orchestration" of existing concepts and theories, with major theoretical import from comparative politics and theoretical policy analysis'.
In this article, we understand and analyze Europeanization in terms of translations and mediations. Following the tradition of Scandinavian institutionalism, the point of departure is that translation is a process characterized by its simultaneous process of movement and transformation (Suarez & Bromley 2015, 145). This means that policies are not merely transferred (in this context, from the EU to the local level) but rather they are changed to something within a new context (see e.g. Czarniawska 2008; Czarniawska and Sevón 2005;Lavén 2008;Clarke et al. 2015;Munday 2016). In other words, translation is a process where new content 'is put togetheror assembledfrom a variety of elements that are always in the process of being reassembled in new, often surprising ways' (Clarke et al. 2015, 9). In the words of Bruno Latour (1986, 267), processes of translation are 'in the hands of people', meaning that people may act in many different ways by 'letting the token drop, modifying it, or deflecting it, or betraying it, or adding to it, or appropriating it'. Re-formulations can entail changes not only in content but also in form (Suarez & Blomley 2015, 146). In this article, translation is conceptualized in terms of both substantive and organizational changes in local policies and politics. In this regard, the EPA studied here act as a policy tool or a device, but in a broader theoretical sense, also as a process of change enacted through actors at regional EU offices, which are in turn produced and reproduced as experts in these enactments. Our point of departure is that it is necessary to consider these processes not as a transfer of knowledge, but rather as processes of translation that are inherently necessary for fitting into the recipients' local contexts (cf. Latour, 1986). In relation to Europeanization, this can be understood as informal bottom-up processes of adapting, accommodating, and adjusting (cf. Alexiadou & Lange, 2015). It is also important to emphasize that processes of translation create not only changes in policies in terms of substance or content but also changes in organizational outcomes (Latour 1986;Lavén 2008, 32;Czarniawska and Joerges, Sevón 1996). According to Czarniawska and Joerges, changes occur when the materialization of ideas becomes concrete (Czarniawska and Joerges, Sevón 1996).
From this follows that intermediary actors, or EU experts, cannot be seen as passive links 'simply diffuse[ing] a fixed set of ideas and practices' between the EU and local governments (cf. Söderholm and Wihlborg 2013, 268). Here, we follow Latour (2005, 39), who defines a mediator as someone or something with the ability to 'transform, translate, distort, and modify the meaning or the elements they are supposed to carry'. This is in contrast to an intermediary, who simply 'transports meaning or force without transformation: defining its inputs is enough to define its outputs'. This study has found that, while EU experts are often described and comprehended by local government employees and Europeanization scholars as 'simply' transmitting information/knowledge the empirical analysis of this article will show that, in practice, these actors are not intermediaries, but rather mediators who act upon ideas which are translated and acted upon in the process. In this regard, the mediators are also produced and re-produced in the translation processes.

Material and methodological approach
The analysis uses Sweden as its empirical case. Sweden is a decentralized unitary state with a Scandinavian tradition of strong local self-government and a relatively high degree of local autonomy (Loughlin 2001, 5). The Scandinavian countries are interesting to analyze from a translation perspective. They are strong welfare states with a decentralized system of welfare provision. In Sweden, municipalities are responsible for the provision of education (as one of the welfare services provided on the local level). The education system was decentralized during the 1990s, providing opportunities for increased local self-government, including substantial influence over organization, financing, and monitoring (Quennerstedt 2012, 57).
The empirical material used in this article consists of 25 EPA analyses from 18 (of 33) municipalities in the region, which were conducted by two separate but similar regional organizations, The SALA and the EU office of Scania North East (SNE). Both SALA and SNE are regional organizations that work for and are governed by several municipalities. Each EPA is a document of approximately 10-20 pages that summarizes the matching process of municipal budgets and visionary documents with EU funding possibilities. They are commissioned and approved by the municipal council, and as such, can be regarded as a guiding municipal document. The empirical material is also supplemented by semi-structured interviews (cf. Edwards and Holland 2013) conducted by us during 2014-2016. In total, we interviewed nine EU coordinators employed in these municipalities (several municipalities in the region are small and do not have an administrator or civil servant working exclusively with EU-related issues). The material consists of interviews with five civil servants employed by either SALA or SNE as well as information available on websites and complementary information material.
The main methodological approach is an interpretative textual analysis which applies Carol Bacchi's 'What's the problem represented to be?' framework (WPA framework) to the 25 EPA analysis. The aim is to understand translations through close readings of how existing municipal goals are re-formulated to 'match' the framework of EU funding. This approach motivates the usage of Bacchi, as core focus in her approach lies on how policies are problematized and thus made governable. The point of departure in Bacchi's framework is the recognition that policies, plans, or proposals are not merely a reaction to set policy problems, but rather policies create and produce problematizations or understandings of the policy problem at hand. These problematizations in turn affect which solutions are deemed possible and/or desirable (Bacchi 2010, 2). In this regard, a distinction is drawn between discourse as a particular way of talking and the understanding of discourse as 'constituting the objects about which it speaks and constraining what can intelligibly be said about them' (Teghtsoonian 2016, 335). In addition to providing a delimited understanding of a policy area, problematizations also entail an understanding of the persons or groups to whom the policies are addressed. In this regard, policies not only reflect subjects but also create subjects such as the 'troubled youth', 'single mother' or 'welfare recipient' (Bacchi 2010, 6). In other words, problematizations create different subject positions (e.g. discriminated women) as well as different policy solutions (e.g. affirmative action, quotas)which means that the way an issue is problematized has very real effects for the groups involved (Bacchi, 2012). This creation of different subject positions is also examined in the empirical analysis, thus highlighting the link between problematizations and subjectification.
In the empirical analysis, we analyzed the EPA analysis in two steps: First, we analyzed how municipalities themselves understand the education problems (problematizations) and then we analyzed how regional administrators/consultants change these problematizations in the EPA process. Potential differences in problem formulations should not be seen as a deliberate form of misrepresentation (cf. Bacchi 2010). Rather, the point is to scrutinize the possible changes that arise as a consequence of problematizing policies in a new 'EU-friendly way'. In light of the overall theoretical framework, these processes of re-problematization are understood as processes of translation from the EU-context to the local context. All EPA documents have been read in their entirety, but an emphasis has been placed on suggested EU-funding opportunities in relation to education. Education is a highly regulated policy field in Sweden and also a mandatory municipal welfare service, which means that Swedish municipalities provide these services within the frame of local self-government. Here, we have chosen to look at education as an empirical example, both because it is a welfare policy area with presumably little direct influence from the EU and because of the potential impact it may have. Hence, the education area is the least likely policy example to undergo change (cf. Seawright and Gerring 2008), meaning that if processes of recompartmentalization are found here, the phenomenon is also likely to be found elsewhere in local government. In terms of methodology, it is impossible to isolate the influence of the structural and investment funds vis-à-vis other sources of international influence on European and Swedish education policies, such as the 1996 OECD report, 'Lifelong learning for All', the OECD's PISA (Programme for International Student Assessments) reports, and the European Commission's European Year of Lifelong Learning programme (cf. Dahlstedt 2009, 77). In this regard, the mediation and translation of funding opportunities is merely one of many forms of influence.
Two sets of questions chosen from the WPA framework are posed to the empirical material in order to scrutinize problematizationsand in our case more specificallypotential changes in the problematizations in the EPA analysis 1 : (1) What is the (education policy) problem represented to be by the municipalities in the EPA analysis? How is the education policy problem re-problematized in order to fit the EU framework? (2) What effects are produced by this representation of the problem? Effects here refer to the subject positions created as well as the possible spaces for action (cf. Ekström 2012) and 'spaces of action' refer to both organizational and substantial changes.
In the following section, the analysis will be situated in the municipal organization, so as to enable the analysis of the organizational changes. Moreover, EPA will be described in some detail before moving to the empirical analysis, where the documents are analyzed.

Contextualizing the study: on the duality of Swedish municipalities
Sweden's 290 municipalities are primarily responsible for mandatory welfare services such as preschools, schools, and care of the elderlya responsibility that financially dominates municipal services (cf. Bäck and Johansson 2010). These areas are regulated through legislation that allows for a certain amount of discretion, which means that differences between municipalities are quite common in terms of both levels of services and their organization. In addition to welfare services, municipalities are, to an increasing extent, also involved in processes of development planning, such as urban renewal, sustainability, and integration (cf. Montin and Granberg 2013). These two broad areas/local responsibilitieswelfare and urban developmentare often conceptualized as mandatory and voluntary services, respectively, and as such, reflect the dual roles of Swedish municipalities. In terms of organization, mandatory policy areas are often sector-divided and subordinated into different committees, for example, the Education Committee (under the Education Act) and the Social Care Committee (under the Social Services Act). In contrast, voluntary policy areas (such as tourism and urban development planning) are in general not subordinated a sector committee, but instead fall under the direct responsibility of the Executive Committee and City Hall. In other words, mandatory and voluntary policy areas differ in terms of legislation and organization (Montin 2010), where the main difference is that voluntary services are: (a) not as regulated as mandatory welfare services and (b) are often implemented in temporary organizations such as projects or partnerships (with or without EU funding). In many ways, the division between the two roles is analytical, but the areas are, as we will show, becoming increasingly interrelated (cf. Montin 2010, 127).
In line with this, local government policies can be said to be 'compartmentalized' into different sector committees and administrations, where the sectors are sometimes described as 'silos' separating municipal policy areas from each other.
Compartmentalizing is a concept often used in psychology to describe a process in which a person separates one area of thought from another to avoid cognitive dissonance or mental discomfort caused by conflicting values, emotions or beliefs. In this article, the re-compartmentalization metaphor indicates a process whereby a policy 'belonging' to one compartment is 'moved' to another. In our empirical interpretation, this includes issues being moved between administrations (i.e. from education policy to social policy) from mandatory to voluntary policy areas or from permanent to temporary (project) organizations. In fact, the ambition of many EU funds, such as the European Social Fund (ESF), is to encourage a process whereby projects shift (after 2-3 years of funding) from being temporary to becoming a permanent organization, and this is integrated in everyday policies. However, we can show how processes are able to move in the other directionfrom the permanent to the temporarya development that has been discussed in the literature on projectification (cf : Fred 2018;Godenhjelm 2016;Lundin et al. 2015;Packendorff and Lindgren 2014;Midler 1995;). 2 When we discuss possible organizational changes that the EPA analysis might suggest, these can be understood in relation to the municipal organization, as described.

EPA
As regional development becomes more integrated with the development of the EU Structural Funds (Johansson 2010, 110ff), economic growth is increasingly facilitated through different types of collaborative bottom-up approaches, where mediation and the specific actors who are constructed in these processes are a response to the external (EU) and internal (regional/local) processes (cf. Suarez & Bromley 2015). As mentioned, we have examined the EU offices of two Swedish regional organizations working with EPA, SALA and SNE. Both are regional collaborative organizations including a number of municipalitiesin the case of SALA, 33 municipalities, and in the case of SNE, nine municipalities. In the interviews, municipal respondents from smaller municipalities explain that it is too difficult to keep up with the developments of EU-related policies, and thus, the responsibility to do so is more or less delegated to organizations such as SALA and SNE. Thus, these mediating EU-expert organizations have great legitimacy within the municipalities.
The explicit purpose of these organizations is to, among other things, support and strengthen the municipalities' EU-related work in different ways. On their website, the SNE states that they are 'a joint resource that works on behalf of the municipalities […]. They are commissioned to enhance municipalities ability to utilize EU funding opportunities in local development work', thus emphasizing that they work on behalf of the municipality. It is further noted that they 'have been able to facilitate the acquisition of 280 million SEK [approx. € 27 million] in different projects' (SNE webpage 2018), 3 indicating that part of the organizations' legitimacy is based in output terms, meaning its success in securing grants and EU funding for its municipalities. In an interview with a local EU coordinator from one of the participating municipalities, the respondent answers the question, 'What does SNE help you with?' with 'They support our work with project applications… they are very helpful and positive and they want to help! I think that they are an important partner, very important!' (Interview, local EU coordinator, October 2016). Similarly, SALA 'aids the municipalities of Scania in regards to their EU-related work […] with the ambition of promoting more municipal actors to take part in the opportunities offered by EU membership', including 'support in project development with focus on EU-funded projects and partnerships' (SALA webpage). 4 This image is also supported in interviews with EU experts on the regional level. One of the experts who also conducts EPA analyses describes his role as a 'developer, someone that creates additionality to ordinary municipal activities' (EU Coordinator employed at a regional authority. March 2016). He continues by emphasizing that [SALA] 'suggests a variety of activities suitable for EU funding, but the municipalities themselves have to decide which activities they want to work with' (interview). In this regard, the interviewees from both the municipalities and the region see this as a straightforward process of information clearly anchored both institutionally, with the development of EU-expert organizations (which have an explicit mediating role, working on behalf of the municipalities) and regionally, with the development of a tool that is described as '100% politically sanctioned' by the regional coordinators, with reference to that it takes political documents as its point of departure. In this regard, regional partnership arrangements open up not only to new types of actors and relations but also to 'new ways of legitimizing political activities' (Johansson 2010, 116-117).
More specifically, EPA was initiated in light of the ambition of a number of local governments in Scania to enhance the ability to 'match' municipal needs to different EU funds and programmes. The tool was borne of a project initiated by SALA and funded by the ESF with the explicit ambition to develop EU competence among key staff in the region of Scania (SALA 2006-08-10; ESF application 2006). In an interview, the origin of the EPA is described in the following way: During the project, we talked a lot about the need to analyse the problems and needs of the municipality before applying for EU-funding as it is only after conducting such analysis that you know what the money is needed for. Then you are ready to write and send in an application […]. (EU coordinator employed at a regional authority. March 2016) EPAs have become popular with municipalities over the past couple years. One possible explanation for its increasing popularity is the previous tendency to initiate EU projects that were possible to attain but not actually sought after in the municipal administration. In interviews with EU coordinators, the problem of 'unwanted' projects is expressed and they emphasized the 'importance of only initiating projects that we actually want' (EU Coordinator of a municipality in Scania, March 2014) including a 'plan for which projects we should initiate' (EU Coordinator of a municipality in Scania, March 2014).
Related to the problem of projects that were not derived from municipal needs is the perceived problem that EU-funded projects lack 'democratic anchorage' (see e.g. Sørensen & Torfing, 2005). Sørensen and Torfing (2005, 201) describe democratic anchorage as something that is 'properly linked to different political constituencies and to a relevant set of democratic norms that are part of the democratic ethos of society'. In relation to EU-funded projects, the critique is often that they are not democratically anchored or 'in line with the popular will expressed by the political majority of the elected assemblies' Sørensen and Torfing (2005, 201). However, in this regard, several respondents described the EPA analysis as a positive experience, as it uses municipal documents as its point of departure. It is also with this backdrop that we can understand the regional EU Coordinator's emphasis on the EPA being '100% politically sanctioned', as mentioned in the preceding quotation.
In the following section, the translation from calls for the EU's structural and investment funds to Swedish local policy within the education field will be discussed using the questions from Bacchi, as presented earlier in the text.

Re-compartmentalizing education
Although it is a national responsibility, education is an increasingly important part of EU cohesion policies, or as noted on the EU website's homepage: EU countries are responsible for their own education and training systems, but the EU helps them set joint goals and share good practices. (https://europa.eu) 5 In this regard, education is an example of how intertwined structural and investment funds are with broader EU agendas of socio-economic development, including EU 2020, 6 and previously, the Lisbon Agenda and the Social OMC, with their governing through common objectives, guidelines and indicators (cf. Daly 2012). This is also noted in several EPA analyses, such as that of the municipality of Trelleborg, which states that: Generally speaking, education is high on the agenda for the Europe (2020) strategy […] within the field of education, there will thus be great possibilities to work with EUfinanced activities. (Trelleborg, 2013. Rough analysis 2013 The renewing and modernizing of education has been conceptualized as part of an ongoing Europeanization of education (Alexiadou 2016;Alexiadou 2014). The efforts are specified in terms of the EU playing a 'key role in supporting and supplementing efforts to improve and modernize EU countries' education and training systems' (webpage). 7 Within the frame of these forms of soft co-ordination, the EU's ability to govern is dependent on its capacity to involve relevant stakeholders in different ways (Sørensen and Triantafillou 2015). In other words, these 'non-coercive processes' are based on the will of local and regional actors to participate and agree on common political objectives (cf. Bruno, Jacquot, and Mandin 2006, 520). In the following, the understanding of education in the EPA analysis will be teased out. Here, we first describe the municipal problematizations, and then we relate them to the suggested funding opportunities. Two main types of 're-problematizations' were found in the material.
Representing the educational policy problem Ifrom poor grades to need for more innovative education and teaching Education in Sweden is currently a policy area that is associated with certain challenges, where two problems often raised are decreasing school results and decreasing horizontal equity (National Agency for Education 2009). It is evident from the EPA analysis that several municipalities in Scania are experiencing challenges within these fields. In the documents analyzed in the EPA processes challenges that are raised in regard to education often include issues such as a need for improving school results and grades, increasing the number of students that are eligible for higher education and decreasing the number of dropouts (see e.g. the municipalities of Hörby 2013; Hässleholm 2014). One example in this context is the municipality of Hörby: […] students who complete primary school should have the possibility to enter higher education. In addition, young people that lack upper secondary education should be given the opportunity to attain basic qualifications necessary for further studies. (Hörby 2013, Detailed analysis 2013 The municipal problematization is lack of incentive or ability to pursue further studies, as well as the need of support for students that lack qualifications for entering higher education. The perceived reason for these inadequacies and local challenges are not included in the policy documents that are cited in the EPA, but in the Swedish context issues of education are in fact highly politicized, where problemsas they are publicly debatedare generally derived to either lack of monitoring (Almer 2014), lack of economic resources and growing inequalities (see e.g. Scocco 2014) or lack of order and discipline in schools (see e.g. Adcock 2016).
In this section, we move from municipal priorities to how these issues are problematized in the EPA analysis. The EPA from the municipality of Hörby (as well as the municipalities of Bromölla (2014) and Ystad (2012)) introduces new problematizations in the way that solutions are presented as opportunities to tackle these problems through EU funding. An example of this is the suggestion that problems in schools should be formulated in terms of 'skills development for the teaching staff' (Hörby 2013, Detailed analysis 2013. To this end, schools can apply for funding from the ESF for projects aiming to improve teachers' performance. Municipalities are also advised to apply for Erasmus+ funding: Erasmus+ strategic partnership school can be useful here as such a project offers the opportunity to work with organizations in other European countries on improving the quality of education […]. (Hörby 2013, Detailed analysis 2013 In Erasmus+, the ambition is that teaching institutions can improve by best practices and learning from other cities in Europe (Hörby 2013, Detailed analysis 2013. This way of soft governing and learning through best practices is significant for EU's governing in the education field, where ambitions, in accordance with the previous quotation from EU's stance on education, are described in terms of setting 'joint goals and shar[ing] good practices' (https:europa.eu). 8 The municipality of Tomelilla is another example of a municipality that has highlighted education in their own prioritizations and plans. Tomelilla (2011) specifies four different aims, which can all be traced to different sections in the Swedish Education Act. The municipal problematizations thus place problems and solutions clearly within the organizational affinity under the Education committee and the permanent organization of local education administration. The four municipal goals are that: [a]ll teachers must have a plan […] based on individual students' needs and conditions in the learning environment; […] every organization (within the school) should have a strategic work against the degrading treatment of people and a plan for the same; […] All children in need of special assistance must be recognized and supported; […] Every organization should systematically monitor and evaluate the results to improve achievement. (Tomelilla 2011, Detailed analysis 2011 In the EPA, these aims are related to issues concerning staff, and the EU mediators suggest opting for ESF projects that promote skills development among teachers, especially if it 'required by a large portion of the staff' (Tomelilla 2011, Detailed analysis 2012. If, on the other hand, the municipality of Tomelilla were to focus more on 'learning from others' the programme 'Interreg IVA Öresund' would be more appropriate (Tomelilla 2011, Detailed analysis 2011. In addition, it is suggested that 'more targeted interventions can be made through international exchange within the Lifelong Learning Programme and Nordplus' (Tomelilla 2011, Detailed analysis 2011, 6-7).
In several cases, education is also further related to issues of employabilityanother prioritized area in EU 2020. The relationship between education, lifelong learning and employability has been critically highlighted in the Swedish and European context (e.g. Zarifis et al. 2014), and the structural funds are seen as a push factor for discourses and practices of employability in several studies from the Swedish context (see e.g. Vesterberg 2016; Dahlstedt 2013) as well as in other European countries (see e.g. Sanchez Salgado 2013; Verschraegen, Vanhercke, and Verpoorten 2011). Within the EU context, employability is described as an ambition to 'tackle youth unemployment by improving young people's skills and employability' (https://europa.eu). 9 The EPA analysis from Örkelljunga states: Erasmus+ is composed of a variety of activities that provide very good opportunities to develop the education area […] Part of the programme focuses on modernizing the area of education and promoting innovation, entrepreneurship and employability. (Örkelljunga, Detailed analysis 2015, 2) As mentioned, the ambition to improve the quality of education is central in EU 2020 and in the quotation education is related to another goalthat of employability. The connection between these re-problematizations and solutions can also be traced back to the EU 2020 indicators, 10 where education and employability are key features. In the European context, these issues are related to those of skills, or as it is developed in the EU 2020 goals to '[…] enhance the performance of education systems and to facilitate the entry of young people to the labor market' (European Commission 2010, 3). Here the simple assertion is that '[b]etter educational levels help employability [...]' (European Commission 2010, 9). It is no surprise that funds are aligned with the overall ambitions stated in EU 2020, and here, the regional actors mediate their knowledge of EU goals through the EPA analysis. In this regard, the EPA analysis can be seen as a mediating tool where policy changes are suggested based on their compatibility with EU goals within, in this case, the field of education. In a broader sense, the above can help us understand how the EU stance on education is translated to the local Swedish context. In the following section, these processes of mediation and translation will be developed further in relation to the goal of employability.
Representing the educational policy problem IIfrom poor grades to social exclusion and 'employability In the EPA analysis, the regional EU mediators in several instances suggest reproblematizations of education in terms of employability. In several municipalities, (e.g. Kävlinge, Bromölla and Hässleholm), the EPA suggests measures related to youth unemployment to handle challenges in the field of education. In the Municipality of Hässleholm, Rough analysis (2014a, 7), the municipality itself formulated problems in terms of 'improving school performance' and 'targeted intervention for children and young people […]' which includes creating 'a positive environment' as well as developing ways and 'tools for including children's perspectives in governance'. In the EPA analysis, these problematizations were related to EU-funding opportunities in the following way: In some cases, collaboration within the European Regional Fund can be considered […]. One can possibly imagine different combinations of collaborations between schools and businesses in order to find personalized training. The focus of such an application must be on the benefit to the business community. (Hässleholm 2014b(Hässleholm , detailed analysis 2014.
Further, in regard to the municipal priority of '[e]arly and coordinated efforts to encourage a positive environment for children growing up', the following is suggested in the EPA: The skills required to develop and create security and a positive upbringing environment can be financed by the three programmes within the ESF.
[…] In this case one example could be Programme 1, which includes employed and non-employed persons and focuses on conversion, skills and the individual's position in the labour market. The link between education and the labour market must be strengthened. (Hässleholm, 2014b, 6) In other words, the problematizations shifted from 'improving school results and educational environment' to 'getting young people employed'. This also entails a shift in subject positions, namely, from 'struggling school children' to 'unemployed adolescents'. Besides this substantial shift in content, the administrative affiliation shifts in these suggestions from the municipal education committee to the temporary realm of collaborative projects within local development and the business sector. In other words, in our interpretation, this entails a re-compartmentalization both in content and in organization.
In addition, the EPA analysis suggests merging different forms of social exclusion in order to enhance funding opportunities and facilitate EU projects. This understanding of social exclusion is again characteristic of Social Europe. Using the ESF as an example, this is described in the following way on the ESF homepage: As well as giving education systems a boost, the ESF is also focusing on their customersschool children, university students, workers and job seekers seeking training and new skills. Many projects are taking steps to reduce early school-leaving and ensure young people have appropriate skills and qualificationsparticularly among disadvantaged groups such as minorities and those with an immigrant background. (ESF homepage). 11 In this way of understanding education as employability, school children become 'customers' who act in the same capacity as 'workers and job-seekers seeking training and new skills'. This also creates new subject positions where 'school children with poor performance' are instead problematized in terms of 'troubled youth'. Another example in this regard is the municipality of Hörby. In the EPA, it is argued that '[s]everal of the priorities tangent each other and can therefore advantageously be combined into larger projects, especially if you want to work within the European social fund' (Hörby, 2013, 6). In this context, the suggestion in the EPA is to merge education with 'preventive social work': It would probably be possible to create a cross sectoral municipal project for skills development for staff who work with children and young adults in different ways. It would also be beneficial to invite other relevant organizations, such as business owners and associations […] (Hörby, 2013, 6-7) The quotation above indicates a shift in problematization from 'lack of eligibility and/ or will to enter higher studies' to 'social exclusion' and 'youth at risk', where the latter is a key priority for EU 2020 and the ESF. Here new subject positions are created, as well as new forms of organizationsinstead of seeing poor grades as an education problem which affects schoolchildren within the realm of education, it is seen as an overall youth problem that needs to be solved in temporary collaborative efforts such as projects. Again, this is an example of how these understandings are translated to the Swedish context via EU mediators.
To summarize, two main problematizations can be derived from the material. Regardless of whether the municipalities themselves perceived the problem as poor results, drop-outs or the decreasing eligibility to enter higher education, the problem was problematized as either as (a) lack of competence and skills development among staff and solved through different programs for increasing skills, including transnational development projects (so-called cross-loading) and best practices or (b) young people experiencing social problems understood in terms of employability and social exclusion, and solved through collaborations and temporary projects. In these processes, problems that were conceived of as belonging to the educational realm in municipal documents were problematized in new ways in order to fit the prioritized municipal points into the EU-funding framework.

Conclusions
The structural funds are inherently and explicitly a multilevel endeavor, and the governance architecture brings with it not only new power relations between levels of policy-making but also new roles (cf. Stephenson 2013, 828). The aim of this study was to enhance our understandings of how European cohesion policy is mediated and translated through local processes of preparing and applying for project funding from the European structural and investment funds. A second aim was to investigate the substantive and organizational changes that occur in these processes of translation. Our analysis focused on the EPA, a process with the explicit ambition of 'fitting' local policy goals and priorities into the framework of EU funding opportunities.
The analysis demonstrates how municipal problematizations of education are 'reproblematized' by civil servants and consultants in regional EU offices in order to 'fit' already existing municipal priorities and goals into the context of EU funding. The result of these processes was the emergence of not only new mediating roles and functions on the regional and local levels but also the formulation of 'new' and more EU-friendly problematizations of perceived policy problems within the frame of local education policies. The new problematizations, in turn, render the education issues at hand to be perceived in new ways, more specifically, as development policies to be dealt with through EU funding in temporary collaborative organizations and projects instead of problems within the education system to be dealt within the existing educational system and the ordinary municipal budget. This re-compartmentalization entails (at least) two substantive and organizational changes: First, the translation from the realm of education policy (the Education Act) to the realm of collaborative development policy, which means a shift in subject position for the citizens involved (e.g. when 'school adolescents with poor grades' become 'unemployed youth'). Second, and related, an organizational shift occurs in the transformation from permanent to temporary project organizations, which in turn, means a shift from mandatory policy areas to voluntary policy areas. In the literature, this last trend has been conceptualized in terms of public sector projectification, where the implementation of European structural funds has been described as an important push factor (Godenhjelm, Lundin, and Sjöblom 2015;Büttner and Leopold 2016;Fred 2018). In line with the results of our study, projectification thus entails new practices in public administration, affecting how different tasks and policies are organized as well as the organization itself. Yet, it is important to emphasize that policies, in this case, within the field of education, are not subjected to direct change through project funding per se but rather via processes of translation and mediation through embedded regional actors or consultants acting as EU experts on behalf of the municipalities. In this regard, the actors conducting the EPA should not be regarded as intermediaries but rather mediators who interpret and reinterpret change instead of simply diffusing a set of fixed ideas and practices.
From a wider perspective, EU projects and the formulation of regional and local EU experts and expertise are voluntary activities, and the policies at hand generally concern areas outside the EU's formal jurisdiction. Yet, policy areas within Social Europe, such as education policies, have gradually developed through shared understandings of concepts such as social inclusion, lifelong learning and social cohesion (cf. Bernhard 2011, 427;). Previous analyses on governing through soft law, OMC or soft co-ordination have indicated a convergence between EU member states through processes of learning and emulation (e.g. Kahn-Nisser 2014). However, these noncoercive modes of governing are dependent on the active and open participation of local and regional actors, where, through participation, actors agree on norms, political objectives and modes of regulation, while still maintaining some form of local integrity (cf. Bruno, Jacquot, and Mandin 2006, 520). As argued by Borrás and Jacobsson (2004, 190-191), soft forms of governing are ways 'to bypass [the subsidiarity principle] by allowing the EU to initiate co-ordinated action in areas where authority rests exclusively with the member states', thereby questioning the 'assumed political neutrality' of these governing processes (Bruno, Jacquot, and Mandin 2006, 519). We further investigate these power relations through the delimitation and conceptualization of Europeanization in relation to structural funding opportunities in particularwith their ongoing processes of mediation and translationin a context where social policies such as education are embedded in national-and wider-EU contexts.

Funding
This work was supported by The Swedish Research Council .