Responsibilizing renovation: governing strategies and resistance in the context of the transformation of Swedish housing policy

This article contributes to the emerging body of literature in the field of urban studies that addresses the classical ‘division of labour’ between analyses of the workings of urban power at the macro- and micro-levels. Our theoretical framework aims to capture how processes of power are exercised in processes of urban restructuring. In the field of gentrification studies there have been calls for theoretical developments based on analyses of various local contexts in which rent gaps may be exploited in similar yet varied ways. We contribute to this discussion through an analysis of governing strategies and protests linked to urban restructuring in the context of the so-called Million Programme in Sweden’s two largest cities. In particular, we address the consequences of public housing companies being forced to operate according to ‘business principles’. Importantly, we demonstrate how advanced liberal government, under the influence of neoliberal ideology, has largely worked through a process of responsibilization. We discern a chain of responsibilization leading from the macro-, via the meso-, to the micro-level – ultimately involving the individual tenant; and highlight how a struggle that we call a politics of responsibility has taken place around each link in the chain.

Responsibilizing renovation: governing strategies and resistance in the context of the transformation of Swedish housing policy Håkan Thörn and Dominika V. Polanska This article contributes to the emerging body of literature in the field of urban studies that addresses the classical 'division of labour' between analyses of the workings of urban power at the macro-and microlevels. Our theoretical framework aims to capture how processes of power are exercised in processes of urban restructuring. In the field of gentrification studies there have been calls for theoretical developments based on analyses of various local contexts in which rent gaps may be exploited in similar yet varied ways. We contribute to this discussion through an analysis of governing strategies and protests linked to urban restructuring in the context of the so-called Million Programme in Sweden's two largest cities. In particular, we address the consequences of public housing companies being forced to operate according to 'business principles'. Importantly, we demonstrate how advanced liberal government, under the influence of neoliberal ideology, has largely worked through a process of responsibilization. We discern a chain of Introduction I n the past decade an emerging body of literature in the field of urban studies has addressed the classical 'division of labour' between analyses of the workings of urban power at the macro-and micro-levels (e.g. Gray 2022; Lees and Robinson 2021;Risager Skaerlund 2022;Tonkiss 2018). This article links up with this literature, addressing the micro-macro link through an analysis of two urban restructuring programmes in Sweden and the resistance to rent increases and planned gentrification that they have provoked. Theoretically, our analysis focuses on how power is exercised and resisted in the context of state-led gentrification and rent gap governance. Drawing on literature that has emphasized how gentrification is no longer limited to innercity areas (e.g. Lees, Shin, and López-Morales 2015) we particularly intend to make a theoretical and empirical contribution to research and theorization on the links between gentrification and territorial stigmatization (Kallin and Slater 2014;Schultz Larsen and Delica 2019;Slater 2017), and to link these concepts to a governmentality approach. In the field of gentrification studies, there have been calls for theoretical developments based on analyses of various local contexts in which rent gaps may be exploited in similar yet varied ways; and there has been some debate on whether the concept of gentrification is universally applicable. Much of this debate has revolved around the extent to which gentrification, as a concept developed in the Global North, is also useful in the Global South (Bernt 2016;Ghertner 2015;Krijnen 2018;López-Morales 2015;Valle 2020). Further, based on their review of studies of gentrification in the countries of Central and Eastern Europe (CCE), Kubeš and Kovács (2020) have argued that gentrification theory in a post-socialist context only partly resembles that of the classical examples based on studies in London and New York. We argue that this is also the case in the context of the (post-)Social Democratic welfare states (Esping-Andersen 1990), such as Denmark (Risager Skaerlund 2022), and perhaps particularly so, in Sweden, where a universalist housing policy was a cornerstone in the construction of the post-war welfare state (Thörn 2012). In fact, a comparison of the analysis presented in this article with Kubeš and Kovács' (2020) portrayal of how state-led gentrification through urban restructuring in the CCE is driven by the interacting levels of the state, indicates that the shape of contemporary gentrification in Sweden is closer to that of Central and Eastern Europe than countries that tend to dominate the literature (the UK, Germany and the US). In both contexts, central, city and district governments have jointly engaged in a rapid dismantling of public housing, by promoting privatization, abolishing or weakening rent regulations and tenants' protection, providing subsidies for implementation of area-based regeneration and certain spatial planning measures.
Regarding territorial stigmatization, Wacquant' s (2007) original formulation overlooked how this process may also involve a transfer of political responsibilities from the central to the local state, as municipalities are commissioned to lead businesses and civil society organizations to 'develop' poor suburbs, referred to by Peck (2012) as 'devolved governance' in austerity urbanism. Further, in their overview of 119 peer-reviewed articles, Schultz Larsen and Delica (2019) concluded that '[c]oncerning politics it is surprising that there are only a few studies that look into the explicit political production of territorial stigmatization' and that 'the idea of analysing territorial stigmatization as something political as such is quite undeveloped ' (2019, 555-6). In order to address this gap while at the same time contributing to strengthening the micro-macro link in this field, we link up with previous work in urban studies that use a Foucauldian governmentality approach (e.g. Dikeç 2007;Ghertner 2015;Gray 2022;Janoschka and Sequera 2016;Künkel and Mayer 2011;Raco and Imrie 2000;Roy 2009;Shlomo 2016;Uitermark 2014). While our analysis of governing strategies is largely along the lines of this literature, our main contribution consists in analysing how various forms of resistance emerge in connection with these strategies. Most significantly, we use the concept of the politics of responsibility to analyse the power techniques through which urban restructuring is enforced and resisted at the meso-and micro-levels in Sweden' s two largest cities, and how these interact with macro-level power logics. We argue that Sweden is a particularly interesting case for a theoretically informed study of contemporary urban power and conflict, considering that the main body of previous research (e.g. Clark and Johnson 2009;Hedin et al. 2012) has pointed to a rapid and deep-reaching neoliberalization of the Swedish model' s universalist housing policy, with significant consequences in terms of increasing social inequalities. We will also demonstrate how this transformation has been dogged by contradiction and conflict.
In relation to research on urban development programmes and rent gaps in areas subjected to territorial stigmatization in the context of the Scandinavian welfare state, our study shares similarities with, and adds to the recent studies by Christophers (2022) and Risager Skaerlund (2022). An important difference, however, is that our study focuses on cases of urban development and rent gap governance without privatization. In this sense it is similar to the study by Uitermark, Duyvendak, and Kleinhans (2007) of 'gentrification as a governmental strategy' in the context of a poor suburb in the Netherlands, but with one important difference. While Uitermark et al. emphasize that their study' s contribution is to demonstrate how state-led gentrification can be led by public housing companies without profit motives, in our case we study public housing companies with profit motives. In Sweden, public housing is organized in the form of municipally owned housing, and in 2011, the Swedish government passed a law stating municipal housing companies must be run according to 'business principles' (Public Municipal Housing Companies Act 2010, 879). This law provided for a profit motive in previously non-profit Swedish municipal housing companies. 1 At the same time the municipal housing companies are still supposed to shoulder a social responsibility, thus making the public housing sector an example of how, in the wake of neoliberal transformations, the Swedish housing market has, in the words of Christophers (2013), become a 'monstrous hybrid' . We will demonstrate how the contradictions that define this monster provide opportunities for resistance and struggles over rental housing.
Based on our original empirical research on governing strategies and protests in connection with the renovation of rental housing in Stockholm and Gothenburg in the 2000s we aim to contribute to a theoretical understanding of how various forms of power are not only exercised, but importantly, how they are also resisted. Focusing on Stockholm and Gothenburg, we analyse two major 'development programmes' led by municipal housing companies -with a particular focus on the renovation processes and how they have been resisted by tenants. These two programmes were initiated in two different phases of the transformation of Swedish housing. In the case of Husby in Stockholm, restructuring began in the late 2000s, i.e. before the 2011 public housing act was introduced, while in the case of Hjällbo in Gothenburg the process started after the law was passed. Our comparison of the two cases involves looking for similarities and differences in the ways that urban restructuring strategies and resistance were played out, with particular attention to the impact of the 2011 public housing act in the latter case. Questions address how power was exercised and resisted in these processes, what governing strategies and counteraction were employed, and how this resistance was dealt with. Empirically, our research builds on, and is also intended to contribute to, previous research on the forms and consequences of Swedish housing policy, particularly the deepening of gentrification and segregation in major Swedish cities. 2 The article is structured as follows: In the following section, we present the theoretical contribution that we developed based on this research, focusing on the concept of the politics of responsibility. We explain how this approach is linked and contributes to the literature on state-led gentrification and the political production of territorial stigmatization. We then provide the context for our study and clarify our empirical contribution to previous research on the transformation of Swedish housing policy in general, and urban restructuring and renovation processes in particular. We then analyse governing strategies and resistance in connection with the overall processes of urban restructuring programmes in Husby and Hjällbo. This is followed by a section in which we zoom in on the micro-level of governance and resistance in the context of tenant consultation processes.
Theorizing power and resistance in state-led gentrification Our theoretical approach: advanced liberal engineering and the politics of responsibility In order to perform an analysis that captures the exercise of different forms of power on a macro-, meso-and a micro-level, and how they interact, we link up with governmentality analysis (e.g. Miller and Rose 2008). In line with this, we conceptualize the transformation of the Swedish welfare state, and its de-and re-regulation of Swedish housing policy as the emergence of advanced liberal government. We use 'advanced liberal' (Larsson, Letell, and Thörn 2012;Miller and Rose 2008) to underscore first, that while contemporary liberal government is heavily influenced by the political philosophy of neoliberalism, some of its key features are inherent in the modern liberal project as such; and second, that it does not necessarily mean less politics, less regulation or less government in the way that neoliberal dogma stipulates. In relation to the Swedish welfare state and its transformation, this approach means emphasizing a certain continuity, rather than a sharp rupture, between the Social Democratic welfare state, known for a mode of governing by extensive regulation often called social engineering, and what we call advanced liberal engineering -a mode of governing defined by de-and re-regulation ultimately intended to support a programme of marketization.
We discern four logics that define advanced liberal government  in the urban context. First, marketization involves de-regulation and re-regulation to support privatization, which in the urban context has supported market rents, the gentrification of urban districts and commercialization of its public spaces (Lees, Slater, and Wyly 2008). Second, partnerships between public, private and civil society organizations perform co-regulation, a process which in an urban context has been conceptualized by David Harvey (1989) as a shift from managerialism to entrepreneurialism. Third, responsibilization involves encouraging self-regulation and active involvement by business, civil society and individuals in political responsibilities previously associated with state agencies (Burchell 1993), as is clearly visible in the urban context in urban restructuring projects and de-centralized planning (Swyngedouw 2005). Fourth, in the context of urban marginality and the production and reproduction of territorial stigmatization, new forms of disciplinary power are introduced (e.g. Dikeç 2007), involving a discourse of urban fear/security and an increasing emphasis on zero tolerance/coercive policing measures (particularly in relation to stigmatized territories), as well as 'soft' coercive disciplinary techniques applied when civil society agents or individuals do not perform responsibilities imposed on them. In the following section we demonstrate how all four of these logics are at work in the urban restructuring/renovation of Million Programme housing.
Our main focus, and our key theoretical contribution, concerns responsibilization. Our analysis focuses on: 1) devolved responsibilization from the macro-to the meso-level, by analysing how municipal housing companies are given the responsibility to lead partnerships with private companies and civil society actors to develop poor Million Programme areas in a way that involves them in tasks that were previously the responsibility of the state. 2) devolved responsibilization from the meso-to the micro-level, by analysing how municipal housing companies transfer responsibilities to the individual tenant, identified as a consumer of housing. 3) the politics of responsibility as a mode of struggle and resistance in the context of urban restructuring programmes and renovation processes.
We also show how de-responsibilization can take place when politicians or housing companies actively avoid taking responsibility, de-politicizing their own position, thus making the workings of power diffuse and abstract. In practice, however, such de-responsibilization shifts the responsibility somewhere else (Thörn 2016).
Territorial stigmatization, state-led gentrification and rent gap governance Our analysis of the politics of responsibility is intended as a contribution to existing research and theorizing on the political production of territorial stigmatization and how this is linked to the concepts of state-led gentrification and rent-gap governance.
Our contribution regarding the role of politics in the processes of gentrification and territorial stigmatization takes its point of departure in two conclusions made in previous research. First, the argument about how the stigmatization of a place may serve both to create a rent gap and as a rationale for 'fixing' it (Gray and Mooney 2011;Kallin and Slater 2014;Slater 2017). In relation to this, we will also demonstrate how the strategies for 'fixing' an area may sustain and even deepen its stigma rather than removing it. Second, as argued by Schultz Larsen and Delica (2019, 557), territorial stigmatization is not an unintended consequence of attempts to deal with the social problems associated with urban marginality, but is 'integral to contemporary forms of neoliberal urban governance' .
Another important point of departure for our analysis of state-led gentrification is Krijnen' s (2018) elaboration on the distinction between rent gap formation/creation and rent gap closure, making the point that the former does not necessarily lead to the latter. Krijnen also stresses that not only the creation, but also the closure of rent gaps 'reflect the outcome of a political struggle' (2018, 441) -but does not elaborate on this important observation. More specifically, we build on Gray' s (2022) further elaboration of Krijnen' s distinction, using a governmentality perspective. Gray introduces the concept of 'state subsidy gap' to refer to 'the economic gap that the state must fill in order to de-risk development' (2022, 75). This refers to processes in which the state provides the conditions to make urban generation attractive for private investment. Particularly in 'economically risky neighbourhoods' such state involvement is necessary. Drawing on Foucault' s understanding of neoliberalism in his governmentality lectures, Gray distinguishes between the 'positive' and the 'negative utility' of government. While the latter refers to the process of state withdrawal that contributes to territorial stigmatization and opens up rent gaps, the former refers to the state' s active efforts to promote development (through subsidies, land giveaways etc.) that may attract private investment and result in rent gap closure. We will demonstrate how the role of the local state in the urban development projects that we study has involved both a positive and a negative utility of government. We thus share Gray' s Foucauldian understanding of how the role of the state should not just be understood in terms of politics and policies but as the exercise of power in a more profound sense. Gray' s theorization, like the broader field of governmentality studies, does not however pay significant attention to how such governing processes involve conflict, resistance, and struggle (Death 2016). We suggest that this approach needs to be further developed by conceptualizing the various power techniques through which urban restructuring and the displacement involved in this process are enforced at the meso-and micro-level. Further, engaging in the study of such conflicts and resistances also reveals how the governing strategies associated with neoliberal governance, as highlighted in both critical urban and governmentality studies, needs to be seen against a context of historically shaped national, regional and local contexts.
This means that the focus of this article, theoretically as well as empirically, is on those governing processes which many of the studies that we have referred to have identified as significant, without exploring this further. These governing practices play a key role in the creation and closure of rent gaps and gentrification, and most relevant to our article, the resistances that they may encounter, as well as the struggles that may result in rent gap closure and gentrification, or the failure of urban development programmes to achieve such goals. Further, we will highlight how these strategies and struggles link the macro-and microlevels, from national government to municipal policies and down to the level of everyday life in the neighbourhood. Here, we also highlight how the function of the intermediary (Scheller and Thörn 2018), in our case a role often played by the Tenants' Union, but also technical experts and consultants (brought in to facilitate dialogue), is crucial to the responsibilization of agents at both the meso-and micro-levels in the sense of performing the function of translating governing processes from one level to another.
The cases and their context: from 'good and affordable housing for all' to 'business principles' In the 2000s, major renovation projects took place in the context of the so-called Million Programme -a million dwellings that were built between 1965 and 1975 and clearly manifested the Swedish welfare state' s efforts to create affordable housing. Today the Million Programme is home to the poorest section of the Swedish population, in peripheral and poor suburbs that are subject to territorial stigmatization. When a series of major renovation projects were initiated in the 2000s, protests were staged against the threat of renovictions, i.e. displacement due to steep rent increases (of between 50-120 percent) that were part and parcel of the renovation programmes (C. Thörn 2020; Polanska and Richard 2019; Pull and Richard 2019;Pull 2020;Sernhede, Thörn, and Thörn 2016;Westin 2011). This resistance, which has also been scaled up to a national campaign against market rents and privatization of public housing, may be seen as part of a renaissance in struggles around public housing that has been visible not only in European countries with a strong public housing tradition, such as the UK, Germany and France, but also in countries beyond Western Europe (Robbins 2020).
Importantly, we place the ongoing renovation of rental housing in Sweden in the context of the transformation of Swedish housing policy connected with the dismantling of the Swedish welfare state, which began in the 1980s (Larsson, Letell, and Thörn 2012). The first phase of this transformation involved the key elements of privatization of public housing, and state withdrawal of favourable loans and subsidies to municipal housing companies. This in turn led to a housing shortage and increasing speculation and financialization in the housing sector, deepening patterns of gentrification and segregation in Swedish cities, and increasing poverty in Million Programme areas (e.g. Clark and Johnson 2009;Hedin et al. 2012). A second phase began in the 2000s, leading to a new law in 2011 redefining the principles according to which municipal housing companies should be run (Public Municipal Housing Companies Act 2010, 879), which opened the way for major rent increases and more generally supported a further marketization of the Swedish housing system (Andersson and Turner 2014;Christophers 2013;Holmqvist and Turner 2014). This second phase also more or less coincided with major renovation plans for the Million Programme (both by private and public housing companies). Importantly, these plans were in many cases integrated into more overarching programmes for urban restructuring of Million Programme areas, led by municipal housing companies in cooperation with private actors and civil society organizations. Such restructuring of Million Programme districts has, in a neo-colonial style, been conceptualized in terms of 'development programmes' -in contrast with the conceptualization of urban restructuring of inner-city areas in terms of 'urban renewal' . In a similar way to the Danish government' s Ghetto List (Risager Skaerlund 2022), the problematization of the areas that are used to justify these programmes has referred to the Swedish Police Authority' s classification of poor suburbs as 'vulnerable' or 'particularly vulnerable' , based on their level of crime. They thus provide good examples of how such political interventions can reproduce or even deepen territorial stigmatization.
In his influential article Christophers (2013) emphasizes a certain continuity in the shifting Swedish housing system as he argues that, rather than being fully neoliberalized, it is a 'monstrous hybrid' of elements of neoliberalization and 'islands of regulation ' (2013, 896). Its 'monstrosity' being that it contributes to intensifying socio-economic inequality. In agreement with a number of Swedish housing scholars he argues that the most important dimension of the post-war housing policy that has survived neoliberal reform is the 'soft rent control' of the use-value system (which is explained further below). Discussing the 2011 public housing act, Christophers argues that a key question is not whether this brings further marketization, but 'how much further' (2013, 893) -which at the time of writing he believed it was too early to answer. A year later, Holmqvist and Turner (2014) came to the same conclusion. Later research addressing this question demonstrates that whatever was left of post-war Swedish housing policy was more or less entirely a local affair, as municipalities, in their capacity as owners of municipal housing companies, land and planning monopoly (though this was in the process of becoming de-centralized), and leaders of urban restructuring projects, still possessed some instruments to govern the housing market. Several research projects have examined the actions of local government and municipal housing companies (Grander 2018;Salonen 2015;Svärd 2016), particularly focusing on a question arising from the 2011 act which stipulated that Swedish municipal housing companies have to function according to 'business-like principles' . What effect does this have on the capacity of local government and municipal housing companies to act according to political (rather than economic) principles and provide 'good and affordable housing for everyone' , in line with the universalist slogan coined by the Social Democratic government when these companies were created to build and administer public housing in the 1940s? This research concluded that the scope for municipal housing companies to manage the state of rental housing in Sweden has been severely restricted but not entirely constrained. In the words of Grander (2018), the shift towards business principles, demands for return on investment, and from state-financed to financialized housing provision (involving a dependence on financial institutions, tools and capital) has thrown the municipal companies into a new state of financialized universalism. Nevertheless, Grander concluded, the municipal housing companies 'have a latent potential ability to counteract housing inequality' (ix). As shown by Salonen (2015) and Svärd (2016), the extent to which this constrained potential could be realized depends on the outcome of a number of struggles at the local level. This is the main focus of our study, and we will show how restructuring programmes, led by municipal housing companies -and particularly the renovations that are a central aspect of these programmes -have been a key field of struggle for interpreting, acting on, shaping the consequences of, and even transforming, the new conditions brought about by the transformation of the Swedish housing policy.
In previous research on renovations of the Million Programme, Mangold and colleagues (2016, 6) identified a 'risk of societal inequity' due to the dramatic rent increases in connection with renovation so far. A report from the Swedish National Board of Housing, Building and Planning (Boverket 2014) concluded that those moving from the areas under renovation are the most socioeconomically weak groups, often moving to much poorer areas, thus deepening segregation in Swedish cities. Lind and colleagues' (2016) study of a renovation strategy by a municipal housing company, which offered tenants an affordable alternative, confirmed that affordability is a key issue for Million Programme tenants. The term 'renoviction' has frequently been used by researchers to designate the specific form of gentrification that has taken place as a result of rent increases in connection with renovation (Baeten et al. 2017;Polanska and Richard 2018;Thörn 2020). Our own previous research (Polanska and Richard 2019;2021;Sernhede, Thörn, and Thörn 2016;Thörn and Thörn 2017) indicates that protests and public debates about the renovations of Million Programme housing have had some effects, both regarding rent increases and the degree of tenant participation. Nevertheless, previous research unanimously shows that: a) renovation poses a major risk of further deepening gentrification, segregation and social inequalities in contemporary Sweden and; b) more research is needed to understand the conditions, variations and outcomes of such processes.
Further, our research addresses Hedin and colleagues' (2012) observation of an increasing unevenness in the ongoing struggles 'for power over the unfolding social geographies of housing' (Hedin et al. 2012, 460); and Holmqvist and Turner' s (2014, 251) observation on tenants' capacity for resistance -that, partly because Swedish housing reforms have been 'complex and apparently piecemeal, they encounter less resistance from the public' and that 'because the effects are on the micro-level, it is less obvious that the changes are politically driven' . Rannila (2021) observed an ongoing invisibilization of housing violence in urban restructuring processes in Sweden, emphasizing the welfare state' s history and the subtle, processual and slow violence with which issues that were previously a public responsibility are dealt with today. While our research partly confirms the argument that resistance has been constrained, we also point to how resistance to the transformation of housing policy has grown since these studies were published, and how it has become central to re-politicizing the issue. In the following sections we examine two important struggles in this regard, Husby in Stockholm and Hjällbo in Gothenburg.
Urban restructuring by municipal housing companies -the cases of Husby and Hjällbo

Husby and the 'Järva Lift'
The upgrade process initiated in Järva in 2007 was led by Svenska Bostäder, the largest public housing company in Stockholm, with over 27,000 apartments. In 2005, the Social Democratic municipal government of Stockholm gave the company the task of developing disadvantaged areas in the city, and in 2007 the company initiated 'Järvalyftet' (the Järva Lift) encompassing 70,000 inhabitants in 6 residential areas. The goal was to 'elevate Järva into an engine of growth, a cultural meeting place, and a place people want to move to -and stay put' (Vision Järva 2020, 9). That the improvement of safety and security was an important part of this overarching goal should be seen in light of the fact that these residential areas had been the objects of territorial stigmatization for decades. The programme was first implemented in Husby, the area we focus on here. Different strategies were included in the development of the district, through new housing construction, conversion of public rental apartments to co-ops, attracting and supporting local businesses, providing cultural and meeting spaces and improving the physical appearance of the old housing stock and the public spaces. New residents were to be attracted by the high standards of the newly constructed housing, renovated apartments in the old stock and by the mix of tenures.
The initial plan presented by the public housing company in 2007 was followed by massive protests by tenants. The organized resistance in Husby led to changes that included quite extensive re-organization of the public housing company, and an agreement was reached that set different levels of renovation, resulting in a lowest level of rent increase of 25 percent, a stop on demolitions, and the requirement that renovation should always build on consultation with tenants. Tenants' protests in 2007 also resulted in new, 'democratic' goals, including the participation of residents in the process of the development programme (Al-khamisi 2015). This involved a contract between Svenska Bostäder and the Tenants' Union that financed and encouraged tenants' local influence (Boverket 2010, 19). This contract is often referred to nationally as an exemplary way of involving tenants and developing local participation. Moreover, in 2009 a consultation process called 'Järvadialogen' (the Järva dialogue) was initiated in the district, inviting residents to discuss the development of their neighbourhoods. However, this provided limited possibilities to influence the outcome, and mainly had the function of legitimizing the restructuring programme (Al-khamisi 2015; Tahvilzadeh and Kings 2018). Tahvilzadeh and Kings (2018, 115) argue that the protests led to 'the establishment of an urban governance regime' that aimed to manage the conflicts arising from the development plans by propagating consultations with residents and cooperation with local actors. The methods of emphasizing participation and cooperation were strategically used to legitimize the development project, by neutralizing conflicts and silencing opposition (cf. Al-khamisi 2015, 160). The demands raised by tenants in their protests to the plan in Husby nevertheless had a ground-breaking effect on the continued work of the housing company, on local civil society and activism in other parts of the city (Lundström 2016; Rosales and Ålund 2018; Tahvilzadeh and Kings 2018).
Acting as an intermediary between the tenants and the housing company, the Tenants' Union played an important part in the consultation process and had an ambivalent role. Initially it succeeded in empowering the participating tenants and halted the process of consultation, as well as assisting the formulation of tenants' demands that re-responsibilized the housing company and, above all, the municipal politicians who were accused of disinvestment in the Järva area. In recruiting local residents who had often been critical of the process to act as intermediaries, it was sometimes criticized for neutralizing conflict, but most often it empowered them to mobilize in activist networks that raised their critical concerns about the development. In one of our interviews, a key functionary in the Tenants' Union who was active throughout the process argued that it was the local government, rather than Svenska Bostäder, that bore the main responsibility for the restructuring process and everything that went wrong in Husby. As a consequence, the Tenants' Union eventually sided with Svenska Bostäder in 'forcing the Stockholm city government to join the Järva dialogue' (Interview A), thus re-responsibilizing municipal politicians and re-politicizing the issue of increasing residential segregation and inequality (Al-khamisi 2015; SVD 2012; Tahvilzadeh and Kings 2018).

Hjällbo and 'the new strategy for urban development'
In August 2019, the board of Framtiden, the parent company of Gothenburg' s four municipal housing companies, decided on a long-term strategy for what it called the city' s 'development areas' . The first area targeted for this strategy was Hjällbo, a residential area with 2,600 dwellings constructed in the late 1960s. A brochure titled 'The Future of Hjällbo' , produced by the municipal housing company Poseidon in 2019 and addressing the residents of Hjällbo, begins with the simple question 'Why?' and goes on to justify the restructuring programme in terms of a problematization of the area that clearly implies a territorial stigma: 'Hjällbo is a neighbourhood with many qualities but also with big problems. A long-term strategy is required for Hjällbo to become just like any other neighbourhood' (Poseidon 2019). The first part of Poseidon' s brochure presents concerns about 'safety' , and the company assures the residents that it will take strong steps 'to increase security in the area, and in this respect, we are cooperating with the police' (Poseidon 2019). It also states that it already has a 'safety team' in place, patrolling the area seven days a week (Poseidon 2019). In addition to safety and security, the main pillars of the restructuring process are the renovation of all houses, creating a 'mix of dwellings' by building new houses (involving public-private partnership) and introducing the conversion of rental flats into co-ops. While this was not mentioned in the brochure, we also learnt that Poseidon adopted a restrictive 'lease policy' , applied to 50 percent of available apartments. This policy selected new tenants based on income (only students on a scholarship or families with at least one member in permanent employment were accepted).
The restructuring programme was presented by Poseidon to the Hjällbo residents at the municipal school in May 2019, with the Tenants' Union present. Approximately 80 residents attended the meeting, many of them asking critical questions that demonstrated concerns over gentrification as a consequence of the dramatic rent increases in connection with the renovation. Our observations of the process that followed over the next seven months however made clear that the role of the Tenant' s Union was ambivalent. On one hand, the Tenants' Union was frequently perceived by residents as playing along with the plans and rules established by the housing company. On the other hand, it took an active role in encouraging the residents to initiate a process of self-organization that eventually led to a major protest in the public square in Hjällbo in December 2019. Speakers during the protest raised concerns over rent increases in connection with the renovation, and not only addressed Poseidon, but also the municipal and national government by placing the restructuring process in a context which, it was argued, was part of an ongoing gentrification of poor suburbs that, contrary to the rhetoric of politicians at local and national level, would deepen segregation in Swedish cities.
We thus observed an emerging conflict over housing in Hjällbo, following a similar logic to that in Husby, beginning with spontaneous resistance against a housing company based on fears of rent increases and displacement, and developing into a political protest that aimed to re-responsibilize politicians at municipal and national level. However, compared to the Husby processes, which began over a decade earlier, Framtiden' s 'new strategy' , as it was 'tested' in Hjällbo, and the tension and conflict that were created in the process, involved not only similarities but also new dimensions that reflected the lessons learnt by actors involved in the struggle over renovations of Million Programme housing in the 2010s, including politicians, municipal housing companies, the Tenants' Union and the tenants themselves.
First, it was obvious that there were residents in Hjällbo that were aware of, and had learnt from, previous struggles occurring in the 2010s. At the May 2019 meeting, the term 'renoviction' (introduced in Sweden in the 2010s to criticize dramatic rent increases and their consequences) was used by a resident, and in the process that followed, an activist from the most wellknown renovation struggle in Gothenburg in the 2010s was invited to a tenant meeting to share experiences. Second, the Tenants' Union had changed strategy as a result of previous experiences of renovation struggles. As representatives explained to us, they had realized that they had previously been 'held hostage' by housing companies, as they had often agreed to take on the task of presenting renovation plans to tenants (Interviews B and C). As a result, Union representatives initially kept a low profile during the May 2019 meeting in Hjällbo (until they were called upon by residents to explain their position). Third, as a result of the protests against dramatic rent increases in connection with the renovations, and the public debate that followed, in 2019 Framtiden decided that their subsidiaries should follow a policy that they called 'cautious renovation' , which included a rule that tenants should have the opportunity to choose a renovation alternative that would result in the lowest possible rent increase. Fourth, the 2011 public housing act put in place new directives to strengthen the influence of tenants. As a result of this the municipal housing companies and the Tenants' Union signed a 'consultation agreement' to give tenants influence over renovation processes. This agreement is discussed further in the next section, which takes a closer look at how consultation has turned out in practice.

Consultation as a politics of responsibility
Following the emerging resistance against renovation, the Tenants' Union drafted a consultation agreement in 2015, and in 2017 a modified version was co-produced with organizations representing both municipal and private housing companies. While consultation agreements have been implemented previously, this one was specifically designed to establish tenant influence over renovation. The agreement states that tenant influence should be exercised through the formation of a tenant 'consultation group' that should enter into dialogue with the housing company before the start of renovation. It assigns the Tenants' Union a role in the process by stating that it may call the tenants to a meeting at which they would be informed about their 'rights and obligations' in connection with the renovation process (Tenants' Union 2017, 6).
We found that the role of the Tenants' Union was rather ambivalent in the process of consultation. This is partly a result of the contradictory position as an intermediary that they are ascribed in the consultation agreement. On the one hand, the Union was not expected to act as the official representative of tenants' interests in the consultation process, and on the other, the fact that it was allowed to have a presence at consultation meetings in practice meant that they often actively did take the position of representing tenants' interests or were expected to do so by the tenants. The ambivalence of the Union' s intermediary position was further deepened by the fact that, according to the Swedish rentsetting model, it had the role of negotiating the rent increase with the housing company. This model, in which the affected tenants cannot expect to participate in the rent-setting procedure, contributed to de-responsibilizing tenants from direct influence on the cost of renovation. This meant that on some occasions, when tenants questioned the model, the Union sided with the housing company in defending the set-up of the renovation process. In the end, both consultation processes that we studied resulted in tenants mobilizing outside of the Tenants' Union, demanding that the housing company and the municipal politicians take on responsibility for previous neglect of maintenance of the housing as well as the area, making the renovation process (especially calculations of costs) more transparent, and insisting that more influence should be given to the tenants. Thus, the consultation process, and the conflicts that arose in the cases that we studied, can be analysed in terms of a politics of responsibility. In the following sections we take a closer look at the logic of this struggle as it has played out in the cases that we have studied.

State of the property: between maintenance and modernization need
We found that in every consultation meeting that we studied the main issue under debate concerned the rent increase following renovations. Further analysis of these struggles requires a brief explanation of the use-value system, and how it has been modified by the new 2011 public housing act. First, rents are set according to a dwelling' s use-value, rather than its exchange (marketbased) value. Second, based on this, the actual rent for a dwelling is set through negotiation between housing companies and the Tenants' Union. Third, until 2011, rents in the private sector had to be based upon rents in the public sector, meaning that the non-profit housing sector determined rents in the private sector. The 2011 public housing act not only removed the non-profit rule for municipal housing companies, it also removed their role in setting the norm for rents in the housing market.
Because of this use-value model, rent increases -as established through annual negotiations -had previously rarely exceeded a few percent. However, the latest wave of renovations revealed a 'loophole' in the use-value system; it allows for significant rent increases if changes to a dwelling increase its 'standard' . In the 2000s, housing companies began to plan for renovation programmes that involved a number of measures that produce such a 'standard increase' , elevating a dwelling' s use-value: e.g. by installing washing machines and underfloor heating, replacing flooring with floor tiles, upgrading the kitchen extractor etc. These measures were legitimized on the grounds of a 'modernization need' . It should also be emphasized that the rent increase made possible by a particular standard increase, say installing underfloor heating, is not based on the actual cost (in many cases that we have studied, the rent increase for a particular measure covers the cost within a year or less), which makes standard increases highly profitable for housing companies. The 2011 public housing act has fuelled housing companies' interest in launching standard increases in two ways. First, since rents in the private sector are no longer based upon rents in the public sector, private housing companies have seen a new opportunity to increase their profits, which has undoubtedly contributed to the increasing interest in buying municipal housing not only in 'attractive locations' , but also in poor areas with a territorial stigma. Second, since (the often under-funded) municipal housing companies are now obliged to work according to business principles, making standard increases is an opportunity to increase profit, often to cover past losses and fund expensive maintenance (such as replacing the plumbing system). Consequently, both public and private housing companies have launched renovation programmes with standard increases that lead to dramatic rent increases.
The most expensive element of renovation programmes launched in the 2000s -replacement of plumbing systems -is however defined in accordance with the use-value system as regular maintenance and does not allow for a rent increase. Considering that maintenance has systematically been neglected in the poor suburbs, a significant aspect of the renovation programmes thus involves compensating for poor maintenance in the past. As the use-value system is in some respects not entirely clear on what measures constitute a standard increase, this led to a struggle between tenants and housing companies over what should count as a 'standard increase' and what should count as 'maintenance' . Ultimately this was a struggle over responsibility for accepting the costs of renovation. While housing companies, by launching a number of 'standard increases' , could make tenants responsible for covering the costs of neglected maintenance in the past, the tenants' counterstrategy was to re-responsibilize the housing company by arguing that there was a 'maintenance debt' that the housing company had to pay for.
This meant that a large part of the consultation processes on renovation that we observed (when they allowed for tenants' voices), was defined by the different and often contrasting claims of the tenants and the housing companies regarding what kind of measures were 'needed' as part of the renovation. The housing company would claim the need for modernization of the standard of buildings and apartments, using technical discourse, while the tenants would claim the need for maintenance and its neglect in the past, thus raising the responsibility of the company to first and foremost maintain the existing standard. The differing claims thus placed the responsibility for costs on different actors. The tenants' claims of neglected maintenance meant the company should be responsible for the costs, while housing companies' claim of a modernization need shifted the responsibility for costs to future rents, according to the improved use-value.
Whatever the group of tenants in the consultation group discussed regarding the standard was undermined when the housing company raised the need for modernization using different techniques to legitimize this claim. Experts were invited to act as intermediaries during the consultation process, among them engineers and architects, to convince tenants by providing 'objective' expertise on the topic. In this process tenants were thus de-responsibilized in the sense that it was assumed they did not have the necessary technical knowledge and therefore were incapable of making informed judgements about the state of the buildings and apartments -a strategy that Stenberg (2018) calls a black box.
One of our informants pointed out that since many tenants are not fluent in the Swedish language, the use of technical terms added significantly to making the consultation meetings hard to comprehend (Interview D). Also, arguments based on established ways of doing things in the context of previous renovations by the housing company were used to convince tenants about the rightfulness of the proposed technical solutions. One way to resist this strategy of de-responsibilization was when tenants, with the help of the Tenants' Union, claimed responsibility by inviting their own experts: 'We used the local competence […] our own networks and friends in the construction business. We started asking odd questions' . (Interview A).

Responsibilization by choice
The consultation process is reduced in scope to questions of design, while the question of rent is taken from the tenants' control. It is designed to responsibilize tenants by making them choose from options on the design of the renovation, at the same time as the rent negotiation is conducted as a separate process from the consultation, after the specific measures for the renovation programme have been decided upon (c.f. Bengtsson and Bohman 2020;Stenberg 2018). This makes it impossible for tenants to make informed decisions during consultations about the consequences of their choices on the extent of the rent increase.
Further, the responsibilization of tenants by giving them choices during consultation processes is achieved individually and collectively. Tenants are expected to make choices on behalf of their neighbours that will be applied collectively, and to make individual choices on their own behalf that will be applied in their own apartments. The individual choices made in the consultation process are, however, frequently linked to collective preferences. The consequences of those individual choices can also be collective. In notes distributed by the housing company after a consultation meeting in 2019 the responsibility of the consultation group was summarized as follows: 'The consultation group can jointly come up with good solutions for the apartments that are other than standard. If, for example, a glazed balcony is installed for all the apartments and included in the rent, the total cost will be lower than if everyone were to choose it as an option. The consultation group must then make sure that it is something that the other residents are interested in' (Notes, meeting 02/12/2019).
In other words, if the choice was made for all the apartments collectively, the cost would be lower than if tenants made the same choice individually, putting great pressure on the tenants in the consultation group to make the 'right' choices. Moreover, the tenants who took part in the consultation process were also given responsibility for informing other tenants about the choices made. One of our interviewees, a tenant playing a key role in a consultation group, reflected on this: It takes way too much time for me as a person to represent two hundred [neighbours] all of a sudden, to run around on eight floors and knock like a whole night. I mean, my face is familiar in my house. Whenever I meet people in the laundry room, I take the opportunity to ask and check the situation. But I cannot run around half of the area and ask people and that we should have meetings here and there, put up notes and copy them myself. [the housing company] does not even want to help us copy … (Interview E).
Few attempts to support tenants in this process were recorded, instead tenants encountered difficulties in accessing buildings where their neighbours lived, in printing information or in translating it into other languages.
Tenants were also responsibilized during consultation in the sense that demands on their behaviour, knowledge and preferences were clearly articulated by the housing company in the design of the consultation and during the process. However, we also observed how the company used strategies to govern what choices were made. In both the collective and individual choices tenants were asked to prioritize collective/objective preferences over individual/selfish choices when selecting the level/extent of the renovation, design, materials, colours and/or extra options for the apartments or the common spaces. During one consultation that we observed, a tenant who did not want to have a glass door installed in the living room was told by the company, 'You talk about yourself, we take into account all tenants in the house' (Notes, meeting 10/02/2020), thus blaming the tenant for being selfish in their preference. The focus of this governing strategy was to persuade tenants to make the 'right' choices, which should prioritize the preferences of future tenants, not the individual choices of a household, even if these choices would make the sitting tenants' lives easier. This strategy gave tenants the responsibility for making a choice that would also appeal to other tenants and affect the subsequent rent increase.
Tenants were thus responsibilized in the sense of being expected to make wise choices, while the format of the consultation was designed and controlled so that the main concern, the cost, was avoided. And when tenants persisted in raising the cost issue, it became apparent to them that they were de-responsibilized in the sense that they could not take part in the rent-setting negotiation. This created a paradoxical struggle over responsibility, a catch 22 situation (c.f. Stenberg 2018, 10). When tenants asked the price of each choice they were supposed to make, the housing company and the Tenants' Union explained that rents can only be set once it is decided what measures the renovation will include, as they will be based on an estimate of how all the measures taken together have affected the total use-value.
We also found that municipal housing companies in Gothenburg were prepared to 'bend' the use-value model in order to fund renovations. This occurred in response to a decision by their parent company, Framtiden, that 20 percent of the apartments should be renovated without rent increase. Representatives of housing companies explained to us that they would finance the renovation of the 20 percent by increasing the rent a bit more on the remaining 80 percent. However, such a practice goes against the principles of the use-value model, which states that the rent for each apartment should be set according to its estimated use-value. Nevertheless, there are no regulations in place to prevent such 'internal economic redistribution' , which makes it likely that a little extra money from the highly profitable 'standard increases' will fund the minority of apartments that get renovations without a rent increasegiven that the rent negotiations are successful from the housing companies' point of view. In practice this means that the responsibility for covering the cost of affordable renovations for a minority of tenants is transferred from the municipal housing company to the majority of the tenants, who get less affordable standard increases. To us, this practice further confirms that the renovation of Million Programme housing, by circumventing or bending the use-value model, has been an arena for interpreting, acting on and shaping the consequences of the shift to an advanced liberal model of Swedish housing policy.
Finally, we found that when tenants asked questions or refused to take part according to the pre-designed consultation model, they could be silenced, reprimanded or shamed by the representatives of the company, and when tenants asked questions that did not relate to the specific topic of the meeting, they could be corrected with comments such as 'This is a non-question in this forum' (Notes, meeting 27/01/2020) or 'There is no forum for that now' (Interview E). In other words, when the exercise of power in the shape of responsibilization of the tenants failed, disciplinary power could be invoked (Polanska and Richard 2021).

Conclusion
In this article, we set out to make a conceptual and empirical contribution to research and theorization on the macro-micro links in processes of state-led gentrification and territorial stigmatization, by addressing what we argue is a gap in previous research, namely conceptualizing how power is exercised and resisted in the governing of urban restructuring processes. We have suggested that critical urban theory centring on concepts such as rent gap, state-led gentrification, territorial stigmatization and urban restructuring could benefit from integrating insights on the exercise of power gained from governmentality studies. Our own contribution in this sense has been to demonstrate how advanced liberal government, under the influence of neoliberal ideology, has largely been played out as a politics of responsibility. Focusing on the transformation of the public housing sector, and how it has been played out in the renovations of the Million Programme, we have discerned and followed a chain of responsibilization leading from the macro-, via the meso-, to the micro-level -from the state to municipal housing companies, to the Tenants' Union, consultants, consultations groups, groups of tenants -and ultimately to the individual tenant. This process has however not always been smooth; instead, the form of struggle that we have termed politics of responsibility has taken place around each link in the chain. Thus, one of the major findings of our empirical study is that the renovations of the Million Programme have become a key field of struggle for interpreting, acting on and shaping the consequences of the transformation of Swedish housing policy.
Our study was also intended as a contribution to the literature that has demonstrated how and why theories of rent gap governance, state-led gentrification and territorial stigmatization need to take varying political contexts into account. We argued that while restructuring programmes in the so-called Million Programme areas have occurred in the context of the transformation of Swedish housing policy, they also involve a certain continuity with the housing policy specific to the Swedish welfare state. We demonstrated that while the rent-setting system based on use-value, pinpointed in previous research as a piece of regulation that survived policy reform (Christophers 2013), previously had the function of protecting tenants from steep rent increases, housing companies have found a way to systematically circumvent this function. In their resistance to this practice, tenants however initiated a struggle over the interpretation of the use-value model, revealing some of its internal contradictions and ambiguities.
In the recent debate, drawing on the distinction between rent gap creation and closure, several scholars have pointed out that the fact that a rent gap opens up does not automatically lead to its closure and to gentrification, and how politics play a significant role in whether closure happens or not (Christophers 2022;Gray 2022;Risager Skaerlund 2022). For example, Christophers (2022) points to several factors that may prevent rent gap closure: housing companies may lack financial muscle; public pressure may either cause private companies to back down in response to the naming and shaming of their brand, or may lead to re-regulation (as in Berlin and Denmark). We would like to add to this discussion that even when rent gaps are closed, they do not necessarily lead to gentrification. Christopher' s Swedish study is a case in point -Blackstone achieved closure in poor Million Programme areas but the areas in question were not gentrified, at least if we understand contemporary gentrification in Shaw' s (2008) terms as 'a generalised middle-class restructuring of place, encompassing the entire transformation from low-status neighbourhoods to upper-middleclass playgrounds' (Shaw 2008;quoted from Slater 2011). Nevertheless, the social consequences were severe. Christophers does not discuss displacement in the article and statistics for this particular case are lacking, although available macro-statistics make it possible to estimate that 25 percent of the tenants were displaced. And those who remained were impoverished and became more dependent on transfers (Boverket 2014), thus making taxpayers subsidize Blackstone' s profit. This article has not primarily focused on how and whether a rent gap is opened and closed in the two areas under study. While we have highlighted how the Swedish government' s list of 'vulnerable areas' , based on crime statistics produced by the police, have contributed to territorial stigmatization and thus rent gap creation, we have primarily engaged in an analysis of the political struggles that occur once a rent gap opens up and a 'development' programme is set in motion. These struggles have been partly successful. Hjällbo is an ongoing case, but the protests that were staged have so far forced the housing company to at least re-strategize and concede to demands to re-negotiate planned rent increases. In Husby, the renovations and restructuring were modest compared to the original plans, and the eventual rent increases were much lower than those originally presented. 'Partly successful' means that the housing company did manage to increase the value of its property (although hardly getting the return that could have been made under the area' s 'highest and best use' , according to Smith' s (1979) definition of rent gap closure), while passing on the bill to tenants and taxpayers. While the extensive roll-back of the Swedish welfare state has brought a reduction in transfers and social security measures, the latter have still made it possible for groups of people facing rent increases to stay put, while being impoverished in the process. In this case we have seen a pauperization rather than a gentrification effect. We have thus seen an effect that we might call 'gentrification failure' . On the other hand, in such cases where urban restructuring programmes based on a particular problematization of the targeted areas have not led to gentrification, they have instead contributed to reproducing, or even deepening, territorial stigmatization. Further, housing companies have in some cases abandoned larger restructuring programmes, instead focusing on slow gentrification through 'vacancy renovation' , meaning that apartments are renovated one by one, as tenants move out, a practice which circumvents collective resistance.
In terms of explaining recent widespread resistance against gentrification through renoviction in Sweden, a crucial factor is that the public housing stock is still significant. The fact that municipal housing companies still have social responsibilities while also having to act according to 'business principles' has provided an opportunity for collective mobilization to engage in a politics of responsibility. Another partial explanation is the long and strong tradition of housing struggle in Sweden, which produced the largest Tenants' Union in the world (measured in relation to the size of the population). Nevertheless, we show that, as a result of a process of de-politicization resulting from its institutionalization and recent policy reforms that undermined its position, the Tenants' Union initially found itself in a contradictory, intermediary position in the recent period of housing struggles. However, anti-renoviction tenant resistance organized outside the Tenants' Union has pushed it to begin redefining its role. In the process, it began to discover that since the legislation and state regulation that had previously supported its capacity to counteract steep rent increases have been seriously weakened, it needs to engage in political mobilization.
To sum up, we found that advanced liberal responsibilization has involved marketization and de-politicization, as the responsibility for fighting segregation is transferred to municipal housing companies, and that this process was significantly deepened with the 2011 public housing act that subsumed these companies into a business logic. Against this background the agenda of tenants in our cases ultimately developed into a struggle for re-politicization, as they questioned the transfer of responsibilities from national and/or local government to municipal housing companies, specifically demanding more tenant influence and careful renovations and, more generally, a re-responsibilizing of the municipality and the state, and in the latter case a re-regulation of the housing market. This article has thus presented yet another case in support of Robbins' (2020) argument that a new agenda for public housing, including affordable rents, secure tenancies, public ownership and democratic control, is urgently needed.

Notes
1 According to the Swedish statistical office 27% of Swedes lived in rental apartments by the end of the 2021. Approximately half of the total rental stock is privately owned, and the other (less) than a half consists of public housing, which is municipally owned. The municipally owned part has traditionally (until 2011) been run according to non-profit principles and has been accessible to all parts of the population (with no upper income limit). 2 Our empirical material consists of 27 interviews with tenants experiencing renovation, representatives of the Tenants' Union, public housing companies' representatives, as well as municipal politicians sitting on the boards of public housing companies and researchers following the process of renovation conducted in 2019-2021 in Gothenburg and Stockholm. Moreover, we have analysed policy documents, participated in consultation processes in both cities in 2019-2020 and used observations, notes and information material disseminated during these processes in the analysis. Our interviews were semi-structured, and we used an interview guide focusing on the topics of 1) the organization of the renovation and consultation process, 2) the role of different actors, and 3) the role of protests. Interviews lasted for 1-2 h and were recorded and transcribed.
Interviewees were asked for informed consent, were informed on how the collected data will be used and stored, and we stressed the participants' right to withdraw from the research project at any time. During the process of transcription, personal information on the participants was omitted. In cases where politicians or housing company representatives were interviewed, there is still a risk that they can be identified, and they were informed about this risk. Our analysis of the material collected in interviews and through observations has been analysed thematically and parts of it have been presented to the participants and their organizations. Our research is part of the research project 'Toward a sustainable Million Programme' (2018-2022), involving several researchers, including another city (Uppsala) as a unit of analysis, as well as further collected data, and workshops with stakeholders to discuss the preliminary results produced within the project, etc. Our project is, as far as we know, the first to analyse renovation processes in different Swedish cities.

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

Funding
This work was supported by Swedish Research Council for Sustainable Development (Formas) grant no. 2018-00191.