Overcoming the limitations to co-production in shrinking cities: insights from Latvia, France, and the Netherlands

ABSTRACT Co-production often appears as a virtue in academic research and planning practice, particularly in shrinking cities, which struggle to develop effective policies and provide sufficient level of public services. In this paper, we argue that the urban shrinkage context imposes significant limitations to co-production practices, that we urge should not be neglected. Drawing on three contrasting projects in Riga (Latvia), Nevers (France), and Heerlen (the Netherlands), we explore existing limitations and consider possibilities to overcome them. The analyses show that the efficacy of co-production practices is impaired due to a reduced level of trust towards public authorities; selective outmigration; weakened social capital; and political prioritizing of private sector interests over civil society. To overcome these, we discuss the role of several tools, namely mediating actors, independent funding, civic empowerment, and the willingness of public authorities to regenerate power relations.


Introduction
In the last twenty years, policy-makers in shrinking cities have been investigating urban strategies that are both effective and feasible, in the face of two key challenges posed by the urban shrinking context: (1) the lack of planning strategies to address shrinkage at the strategic and policy level; and (2) tightening human and financial resources to maintain infrastructure and public services (Martinez-Fernandez et al. 2015;Sousa and Pinho 2015;Wiechmann and Bontje 2015).In this regard, co-production has been proposed as a well-suited approach for shrinking municipalities.The involvement of civil society is said to provide an opportunity to deal with the manifestations of urban shrinkage and to develop innovative policies relying on local assets and capabilities, which together can lead to alternative paths of development (Schlappa 2017;Matyushkina 2021;Matoga 2022).Drawing on urban shrinkage scholarship that identifies specific characteristics of shrinking cities, which are likely to hinder the implementation and effectiveness of coproduction initiatives, this paper diverges from a perspective that, we argue, might be overly optimistic.Specifically, we point to a tendency in co-production scholarship to assume beforehand the suitability and effectiveness of co-production in shrinking cities, and to the necessity of a more critical approach.
Against this background, we ask what are the limitations to co-production in the context of urban shrinkage, and what could be the possible avenues to overcome them?To answer this question, we draw on cases of co-production initiatives in three shrinking cities, Riga (Latvia), Nevers (France), and Heerlen (the Netherlands).All three cities have experienced long-lasting urban shrinkage, but are different in terms of the strength of civil society and tradition of civic engagement in policy and planning.
Our contribution to the co-production scholarship is two-fold.First, from an empirical standpoint, we argue that co-production as a planning and governance tool should not be idealized because of the limitations of such initiatives in the context of urban shrinkage, in particular regarding weakened social capital, diminished empowerment of disadvantaged population groups, impaired trust relation with policy-makers, and the policy-maker's tendency to prioritize private sector interests.Secondly, from a theoretical standpoint, we move away from the dominant understanding of co-production as purely co-delivery of public services, widely used in public management field.We argue for an understanding of co-production that takes into account, not simply the co-delivery of services, but also the reallocation of decision power to civil society.Drawing on our empirical findings and on previous reconceptualization of co-production focusing on the distribution of power (Mitlin 2008;Watson 2014) we suggest that only when the two conditions are metco-production of decision-making and service provisionmay co-production practices deliver alternative strategies for shrinking cities that stimulate social transformations and address shrinkage related problems.

Co-production in shrinking cities
2.1.Why co-producing in shrinking cities?
It has been argued that engaging citizens in co-production can benefit shrinking cities at two levels: service provision and policy-making (Schlappa 2017).At the level of service provision, governments in shrinking cities struggle with budget deficits because of decreasing tax revenues.Moreover, they often have difficulties attracting private investments, external public funding, and a sufficient number of qualified civil servants.In this context, the participation of civil society through co-production is considered a way to access new resources at less expense (Ostrom 1996).Several studies on shrinking cities share a common assumption that citizens have endogenous resources that they can use to achieve more sustainable results to counteract shrinkage, compared to government-led and top-down solutions (Schlappa 2017;Hospers 2013;Leetmaa et al. 2015).
At the level of strategy-making, local authorities often lack efficient planning tools specifically designed to address urban shrinkage.Moreover, they often lack comprehensive knowledge about local problems and available resources to develop strategic tools with local sensitivity.Historically, planning tools have been persistently designed to accommodate urban growth, which have proven to be inefficient in shrinking cities (Sousa and Pinho 2015;Bernt 2016;Camarda, Rotondo, and Selicato 2015).This increases the demand for the so-called alternative strategies.However, public authorities in shrinking cities often lack the capacity to experiment and innovate under the conditions of budget deficits, rising amounts of obsolete infrastructure, social problems, and bureaucratic rigidity.In this context, grassroots solutions produced through civic involvement are deemed more desirable.This is where the need for civic actors to co-produce alternative planning and governance strategies for shrinking cities becomes evident.
These two directions outline our focus on: (1) co-production of public services defined as citizen participation in public service provision; and (2) co-production of governance defined as citizen participation at the strategic level in policy and decision-making.

Co-production as power reallocation initiative
Many scholars in the field of urban planning use the public management approach to coproduction.It is driven by the rationale for cost-efficient public service provision and often refers to co-production as a process of engaging non-traditional actors 'into which any ideological content can be poured' (McMillan, Spronk, and Caswell 2014, 202).For example, in one of the most commonly used definitions by Bovaird, co-production is 'the provision of services through regular, long-term relationships between professionalized service providers (in any sector) and service users or other members of the community, where all parties make substantial resource contributions' (Bovaird 2007, 847).Some scholars note that, by defining co-production through the organizational form of stakeholder involvement, research overlooks the role of politics and decisional power in configuring co-production processes (Mitlin 2008;Watson 2014;McMillan, Spronk, and Caswell 2014;Turnhout et al. 2019).According to them, co-production has the potential to achieve social transformation only when it aims to empower civil society and produce more inclusive and just planning models in a broader context.In this paper, we aim to take into account the power dynamics in shrinking urban context and explore whether it can improve access to planning and governance processes for groups that lack decision-making power.
Moreover, literature on co-production of urban governance criticizes top-down implementations of participatory processes, arguing that they constitute a threat to democratic accountability and can institutionalize inequalities (Lemanski 2017).This scholarship underlines instead 'activities by civil-society actors both inside and outside of formalized institutions' (Galuszka 2019, 156).Actions that take place beyond legal frames and institutional settings are considered crucial for transformative change of governing practices.In this regard, participatory efforts of public authorities in shrinking cities could advance the critical state and civil societies' democratic perspectives.
In line with Albrechts (2012), who highlights the processual and transformative character of co-production instead of it being a perfect ideal that can be achieved completely, we argue that co-production is an inclusive process of citizen participation, where citizens have sufficient power and resources to participate throughout both the decisionmaking and implementation phases.Only then is co-production able to deliver alternative strategies suitable for shrinking cities and to address social justice issues such as equity, diversity, and democracy (Fainstein 2010).We define co-production in shrinking cities as the engagement of civil society in policy-making and governance, through a reallocation of decisional power, to enable participation in public service provision in such a way that it benefits local residents.Benefitting local residents forms a central aspect.Here, we draw on urban shrinkage scholarship that criticizes growth-oriented strategies as not benefitting local communities and argues instead for strategies aiming to improve the quality of life.
To sum up, our focus falls on exploring the possibilities of co-production in shrinking cities, in terms of its value for delivering municipal services and tools for 'empowered participatory governance' (Fung and Wright 2003).Understood critically as a power reallocation initiative, co-production can deliver the innovative, socially just strategies, which shrinking cities are looking for.However, it is important not to idealize co-production as a planning and governance tool.Therefore, in this paper, we aim to acknowledge the limitations of co-production in order to understand how they can be overcome., Sicilia, and Howlett (2018) emphasize that co-production has been idealized among academics and policy-makers.It often appears as a magic concept, and academic discourse lacks discussion about the potential threats, challenges, and limitations to and of it.That is why we find important to analyse the limitations to the potential of co-production, because of the limits of what civil society and public authorities are capable of in the context of shrinking cities.

Sorrentino
The first limitation to co-production in shrinking cities concerns the impaired trust of civil society in local authorities.This phenomenon has been empirically documented in the cities of Heerlen, in the Netherlands, Blaenau Gwent, in Wales, and Flint, in the U.S.A. (Morckel and Terzano 2019;Ročak, Hospers, and Reverda 2016a;Ročak 2019).The low level of trust can be explained by the lack of transparency in the governments' way of work, complicated bureaucratic procedures, and citizens feeling they have little power to affect decision-making.Besides, it is often a long-term outcome of the perceived incapacity of local policymakers to prevent events leading to shrinkage, such as company and public service closure, which politicians are held responsible for allowing happening.The low level of trust is a governance challenge, because any actions that city officials try to implement will not gain social acceptance, including co-production projects.
Second, the changing social composition in shrinking cities presents a challenge for the development of co-production practices.Shrinking cities face selective outmigration, meaning that residents who are younger or with higher education degrees tend to move out (Fol 2012;Hospers 2014).For this reason, the social composition in shrinking cities is often characterized by a decreasing proportion of middle-class households and an increasing proportion of low-income households (Fol 2012).Whereas the level of citizen involvement in co-production is influenced by variables such as age, socio-economic and immigrant status, and educational level (often associated with organizational and entrepreneurial skills, and political and administrative literacy rates) (Brandsen, Steen, and Verschuere 2018;Thijssen and Van Dooren 2016).For example, Thijssen and Van Dooren (2016) showed that city neighbourhoods with an 'active middle-class' have more active participation, including co-production projects.
The third limitation to co-production concerns the weakened social capital in shrinking cities, understood here as 'the sum of resources, actual or virtual, that accrue to an individual or group by virtue of possessing a durable network of more or less institutional relationships of mutual acquaintance and recognition' (Bourdieu and Wacquant 1992, 119).Studies show that social capital affects residents' capacity to participate in urban processes (Brandsen, Steen, and Verschuere 2018;Thijssen and Van Dooren 2016;Voorberg, Bekkers, and Tummers 2015).In a community with a strong level of social capital, people are more eager to engage with each other and share the resources to co-produce communal goods or services.Moreover, a community having strong social capital can lead to a spill-over effect, in which weakly connected residents benefit from living in a community with a strong network (Thijssen and Van Dooren 2016).In shrinking cities, population decline results in amputated social networks and weakened social capital, as consequences of selective outmigration (Ročak 2019;Hospers 2014;Audirac, Sylvie, and Martinez-Fernandez 2010).Weakened social capital results in a diminished ability of residents to conduct civic actions (Ročak, Hospers, and Reverda 2016b;Le Borgne 2023).
Last but not least, co-production in shrinking cities can be limited by the dominance of private sector in planning processes over the voices of civil society.Existing studies report on successful co-production projects where public, private, and civil society organizations are engaged, as well as citizens (e.g. in Altena, Germany (Schlappa 2017)).However, under the stress of limited economic resources and the intense need to mobilize them, local governments in shrinking cities tend to favour private interests, while drawing the decision-making power away from civil society.As a result, little possibility for equalizing power asymmetries is allowed.Some cases show how public-private collaborations are unstable, short-term, and result in artificial co-production.This has especially been the case in post-socialist urban contexts, but also occurs in other environments (Bosák et al. 2020;Liebmann and Fröhlich 2009;Nedučin, Krklješ, and Gajić 2019;Rink et al. 2012).For example, in his study on the German urban renewal program Stadtumbau Ost [Urban Development East], Bernt (2009) documented the so-called 'grant coalitions' between public actors and private housing corporations, which were formed to satisfy the criteria for receiving a national subsidy and split up shortly after.According to our definition of co-production, the dominance of collaboration by private actors who reproduce or intensify 'growth-oriented' urban politics does not lead either to socially just power balance or to inclusive innovative strategies for shrinking cities.

Qualitative comparative case study
This research was conducted in a qualitative, comparative manner.It dealt with three shrinking cities in different national contexts: the French town of Nevers, the Dutch town of Heerlen, and the Latvian capital, Riga.The research was problem driven (Flyvbjerg 2006) and searched for a deeper understanding of the phenomenon of co-production in the context of urban shrinkage.Case study research works particularly well when the goal of the researcher is to bring to the front a rich understanding of events, experiences of actors, and conclusions that can be deducted from the case.Therefore, in-depth case study research was chosen as the appropriate methodology to grasp the very nuanced and complex manifestations of the caseco-productionwithin the three contexts.The emphasis was put on finding out what are the limitations to co-production in the context of urban shrinkage, and what could be the possible avenues to overcome them.

Case study selection
Empirical studies on co-production illustrate that this phenomenon can be found in diverse contexts.However, some studies suggest that co-production is more likely to emerge and succeed in contexts with a strong civil society and long tradition of civic engagement in policy and planning (Torfing, Sørensen, and Røiseland 2019).Therefore, these two variables are selected for case comparison based on the 'most different' research logic.We looked for case studies with maximum heterogeneity with regards to these two indicators, but high similarity in the context of long-lasting urban shrinkage.This allowed us to examine the possibilities for co-production practices to emerge in shrinking cities despite differences in tradition of civic engagement and strength of civic society.
We evaluated the 'strength of civil society' based on the rate of active citizenship 1 and participation in formal and informal volunteering activities. 2 These data are available through the survey on social participation in the European Union (EU) (Eurostat Social Participation and Integration Statistics in the European Union (EU) 2021), although the most recent data is for 2015.The Netherlands had the highest level of participation in formal and informal volunteering activities among all EU member states (40.2% and 82.3%, respectively) and one of the highest active citizenship rates, with 25.2%.France had an average level on both indicators, with 23.1% and 23.2%, respectively, and a high rate of active citizenship, with 24.7%.While Latvia showed low levels of active citizenship and participation in formal volunteering activities, with only 5.6% and 7.3%, respectively, it had an average level of participation in informal activities (28.3%).Such a difference between participation in formal and informal volunteering activities can be attributed to cultural specificity; that is, Latvian society favours nonorganized voluntary activities more strongly.In order to rank these countries based on the three indicators, we calculated the mean.Therefore, we assume that the Netherlands has the strongest civic society, followed by France and Latvia (Table 1).
Concerning the tradition of civic participation in policy and planning, Latvia has the shortest tradition, beginning from 1994 (Akmentiņa 2020), while the Netherlands and France have implemented civic participation since the 1960s.The Netherlands stand out due to an event that transformed the participation culture immensely.The King's speech of 2013 announced the participation society and the end of the welfare state in its known form.From then on, citizens were asked to actively take responsibility for shaping their neighbourhoods and fostering social wellbeing by taking over former governmental tasks.Therefore, we assume that the Netherlands has the highest level of civic participation in planning and policy, followed by France and Latvia.Each of the three authors of this paper individually approached one of the three cases.For data collection, analyses of relevant documents, qualitative semi-structured interviews, and participant observations were the main sources.In-depth interviews were conducted in 2020 and 2021, in a semi-structured format.The participants were recruited via snowball and network sampling.In Riga, 8 interviews were included in this study.The participants included civil servants from the strategic urban planning and property development department of the city council, members of the NGO Free Riga, a private investor, and artists who were also the residents of the creative quarter.In Nevers, the research included 23 research participants with 17 users of the community centre, 4 staff members and 2 external actors.With them, 15 in-depth interviews were conducted.In Heerlen, 7 interviews were conducted with the actors from the local administration, civil society, and research, as well as the intermediary actor.The researchers provided participants with the consent form with information about the project's objectives, participants' rights, as well as data use, sharing and storage, and received written (or in case of online interviews oral) consents.
The topics discussed in the interviews were similar over the three case studies.First, we were interested in exploring experiences of urban shrinkage through biographical data of individual participants or organizations, and the reasons for acting in a co-production project.Second, we studied the governance processes in implemented projects, focusing on understanding the actor networks, internal power relations, the outcomes of co-production, and the limitations to it.

Overcoming the limitations to co-production in shrinking cities
In the remainder of this paper, we present an in-depth analysis of the co-production projects in the three shrinking cities: (1) the artist-led regeneration of abandoned territories in Riga; (2) the mutual aid system to cope with social isolation in Nevers; and (3) the search for a participatory governance model to fight physical deterioration in Heerlen.

Artist-led regeneration of abandoned territories in Riga, Latvia
Riga was one of the largest industrial centres during the Soviet occupation of Latvia, but after the fall of the Soviet Union, the city experienced a strong outmigration wave and economic decline, followed by the abandonment of residential properties and industrial areas.From 2007 to 2009, Latvia was severely affected by the economic crisis, thus population numbers and the built environment degraded further.In Riga, urban shrinkage is concentrated in the urban core while the city's periphery is still growing.The Riga City Council (RCC) developed several 'carrot and stick' instruments to deal with abandoned properties, including co-financing schemes or tax penalties, but they served as weak stimuli for physical maintenance.Moreover, redevelopments that aim revive not only physical, but also the social fabric of the abandoned territories, were critically lacking.
A turning point happened during the preparations for the European Capital of Culture (ECoC), the status of which Riga received in 2014.The pressing need to find affordable spaces for performances and exhibitions of artistic projects appeared together with the realization of the ECoC team that there was an excessive number of vacant properties in the city.'How come that nearly every 5th building in the city centre is vacant or abandoned, but there is no space for the many cultural, arts and social initiatives?' (Rubenis 2022, 20).This frustration and paradox created the momentum for the foundation of a new artist-led temporary use movement 'Free Riga'.
The leaders of Free Riga had an idea to develop the temporary-use model for the city, which would ensure a systemic change in redevelopment of degraded spaces and give the creative community access to affordable spaces in the city centre.They started negotiations with the Property Department of RCC to turn the movement into an institution.Free Riga registered as a 'public benefit NGO' to establish closer cooperation with the municipality and be able to use municipal spaces free of charge.Moreover, regulations at that time allowed the NGOs to get a 90% tax reduction for properties in their use, which became a significant benefit for private owners to collaborate with Free Riga.Eventually, the RCC 'recognise [d] Free Riga as a valuable organization in dealing with the promotion of creative and social temporary use of vacant buildings' (Rubenis 2022, 9) and signed a memorandum contract, which allowed the NGO to gain access to information on property ownership.
In 2014, Free Riga started their first project, called Tallinn's Street Quarter.The idea was to transform an abandoned courtyard with buildings of former garages into affordable spaces for artists and creative youth to work, perform, and build a community.Initially, the space belonged to the municipality, thus Free Riga asked RCC to use it free of charge.However, the municipality found private investors, who purchased the space, and advised Free Riga to continue negotiations with them.Inspired by the success of the creative quarters 3 in Moscow and Berlin and motivated by the 90% tax relief, the private owner was eager to collaborate with Free Riga to develop a commercially successful project.Later on in the interview with the representative of the property department we found that according to Latvia's privatization policy, the city has an obligation to sell everything that is not used for its direct functions: 'that's why we're trying to get rid of everything.And the list of temporary use properties in Riga is very short; it is so bad that no one has interest in it' (Riga, Interview 2020).
The withdrawal of the municipality in this project resulted in the situation where the NGO had little power to negotiate with the private investor over the direction of the quarter's development.More and more commercial cafes, restaurants and bars came in.The artists felt displaced: 'we became not a high priority for the kvartal [quarter in Latvian], and we feel that.It's not a cultural space anymore; it's a space that makes money' (Riga, Interview 2020).The private owner shared in the interview that the space was not commercially successful either: the maintenance of the large derelict industrial sites requires large and regular investments, and the 'social aspects will die with the lack of electricity, water and trash collection, it all costs money' (Riga, Interview 2020).They added: 'It's like a suitcase without a handleit's hard to carry, and it's a pity to drop it' (ibid).
Free Riga attempted to contribute to the co-production of a governance model for abandoned territories by developing an innovative temporary-use model in response to the municipal problem of physical degradation.However, it failed to durably provide power to the civil society organization, which represented artists and creative youth.Even the memorandum signed between the city council and Free Riga only granted their access to information, but did not allow any structural governance changes.From the beginning, Free Riga created little connection to the wider neighbourhood community, but promoted the interests of a narrow interest groupartists, who were later displaced by the commercial actors.The project therefore did not succeed to deliver long-lasting social value through its attempt to co-production.

A mutual aid system to cope with social isolation in Nevers, France
Nevers is a mid-sized city with 33,279 inhabitants located in the centre of France.Since its population peaked in the mid-1970s, the city lost 27% of its residents by 2018. 4 Historically, the city concentrated administrative and commercial functions for the surrounding rural and industrial region, which suffered several waves of industrial decline.As a result, Nevers encountered rising unemployment, residential vacancy, and poverty rates (respectively 20.3%, 17%, and 23% in 2018). 5Moreover, as is often the case in medium-sized French shrinking cities, Nevers faces population ageing, with a third of the municipality's residents aged over 60.
The case discussed here, called 'Together we go further', was a project to establish a mutual aid system among the members of a community centre, which was the project initiator.In France, community centres provide social, cultural, educational, and leisure activities, with the aim of fostering social, family, and intergenerational ties.Importantly, they help inhabitants to create and implement their own projects, by providing them with human and financial resources.The community centre in Nevers is primarily financed by the national administration, which allocates family allowances (Caisse des Allocations Familiales, CAF), and partially by the local municipality.By financing community centres, national and local decision makers delegate to them not only the part of the social policy, but also the function of civic participation and community building.The aim of the founding partners is that projects led by community centres should be initiated by residents with the help and support of their staff members.Therefore, the community centre is included in the co-production as a middle actor that professionally provides public services together with local volunteers.
Youth outmigration is one of the main consequences of urban shrinkage and often leads to the social isolation of elderly adults.In Nevers, however, where shrinkage is not acknowledged as a comprehensive process, this problem was not considered a political priority by the municipality, especially in comparison to other imperatives such as economic revitalization.In the context of increased isolation caused by the Covid-19 pandemic, the community centre's staff has recognized the negative psychological consequences of forced isolation.To tackle this issue, they came up with a project that would compensate for: (1) the lack of a strong social network for individuals who are socially isolated, and (2) the lack of personal assistance for elderly and/or isolated people.
In the first phase of the project, a discussion group was organized where participants shared the challenges of the lockdown and tried to come up with coping strategies together with the centre's staff members.According to the participants, sharing such experiences gave them good practical solutions to handle isolation and a good feeling of preparedness for future periods of social and mobility restrictions.However, the last discussion group session, which was dedicated to identifying coping tools and resources, became more arduous, as the participants struggled to imagine collective solutions based on their previous discussions, and to feel legitimate to suggest ideas.The lack of empowerment was later confirmed by the community centre's director, who admitted that it was often hard to find proactive volunteers to initiate community projects.
To overcome the lack of ideas from participants, the staff members took a greater initiative in the creation of the project, resorting to a top-down approach, and proposed the solutions themselves.A project was implemented by Jeanne (the real name was changed for confidentiality reasons), one of the centre's employees, who created a mutual aid board, where the participants wrote something they would offer to others and something with which they needed help.She wrote, as an inspiring example, 'I offer knitting lessons' and 'I am looking for someone to take care of my cat'.The members had to ask Jeanne before putting anything on the board, to ensure that all services remained free.She was thus supposed to be the intermediary between the members and the board, as well as between the members themselves.
The interviews with members of the community centre revealed the low level of participant involvement in the co-production of the mutual aid system and a low level of ownership.Most members found it to be a good idea but did not know how they could contribute or participate. 6Additionally, the participants were often already involved in similar mutual support configurations, for instance helping each other with groceries or driving to doctors' appointments.However, emotional involvement through friendship ties was a prerequisite to enable such mutual aid relationships, and, most particularly, trust was necessary for people to ask others for help.As a result, the board was hardly used after its launch.It failed to address the existing need and did not create the ideal conditions to enable mutual aid relationships.The main obstacle to its success was the failure to retain the bottom-up approach throughout the whole project and to sufficiently empower participants to create their own solution.
The main positive outcome of the project was the social interactions fostered throughout its implementation phase.

The search for a participatory governance model to fight physical deterioration in Heerlen, the Netherlands
Heerlen is located in the southeast of the Netherlands, close to the German and Belgian borders.In the twentieth century, industrialization triggered rapid economic and population growth in the city.However, the closure of the mining industry in 1965 brought economic decline and outmigration, which resulted in the long-term urban shrinkage, with unemployment, poverty, and high vacancy among the main symptoms (Ročak, Hospers, and Reverda 2016a).In 2022, almost 87,000 people inhabited the city (CBS Statline 2022).
The area around the North of Heerlen -Heerlen Noordwas the most affected because it used to be the centre of the mining area.With many buildings torn down, open spaces emerged and the whole area needed to be restructured.According to the Dutch legislation, area restructuring as a planning activity is a top-down responsibility.However, due to the lack of finances, the municipality realized they could not provide a new area plan to accommodate the situation.The involvement of citizens emerged as an affordable strategy to deal with the empty spaces.
The municipality developed a policy titled 'Atlas Gebrookerbos', which prioritized citizen involvement in the area regeneration.This strategy profoundly changed the way of work within the municipality: new positions of account managers and an independent broker were appointed, the existing master planning method was abolished, and the formal bureaucratic procedures that obstructed residents' experimentation were refined.
Bureaucracy and long social distances within the municipality are common obstacles when it comes to civic involvement.Many citizens hesitate to realize their ideas, because they know there will be at least several people that they will be referred to, several forms they would need to fill out, and sometimes even years of preparation and collecting all necessary documents.With a good understanding of these problems, the municipality had to change its internal ways of operating.This meant not only letting go of the master planning, but profoundly changing the bureaucratic process.For the latter, the role of account managers was implemented and each civic initiative was assigned one account manager, who would guide them through the whole process.They would be the only contact point in the municipality that the citizens had to deal with, which resulted in increased accessibility to the planning process.
The role of the broker was a profound step in the process.In his interview, the broker described the role as being a 'pacemaker' who helped people develop their ideas on paper, realize them on the ground, or guide them through the formal municipal procedures.The main idea of appointing the broker was to overcome scepticism and distrust of residents towards the municipality.That is why the broker had to be an independent person and not be affiliated with the municipality.The broker was employed at the independent research institute NEIMED.
In 2021, from 78 initiatives altogether, 32 have been realized or were ongoing, and 36 were not continued or were realized elsewhere.Ten of the initiatives were in the stage of idea formation.All initiatives fall under one of the three designated categories: nature, network, or urban farming.The first puts the focus on giving areas back to nature, in the form of reforesting or designing parks.Initiatives that fall under the network category typically have a goal of contributing to tourism or recreation.Such initiatives include bed and breakfasts, urban community parks, and dog schools.The third categoryurban farmingputs the focus on sustainability, knowledge, and education.Examples in Gebrookerbos include urban farms or vegetable gardens, with the focus on agriculture and forestry knowledge.
When talking about Gebrookerbos, it is important to distinguish between the process and the project.Although it was set up as a project that received funding from IBA 7 and had a start and end date, Gebrookerbos was about creating a process: a new method of co-producing with citizens.At the beginning, municipal employees showed resistance as it was difficult to overcome rigid long-established methods.However, even after the project officially ended in 2020, the method of work is still ongoing.Two follow-up projects show the legacy and sustainability of Gebrookerbos.The first one, 'Stadslab Heerlen', is an urban laboratory for creatives and artists who, with the help of the broker, can find vacant spaces for their initiatives.The second, 'N-Power', is an international project between Belgium, Germany, and the Netherlands to generate innovative neighbourhood development strategies with the focus on civic empowerment and mutual knowledge exchange between public sector employees.In their interviews, the project managers shared that money has not been an issue at any point during the project: 'We started Gebrookerbos funding and there was 200,000 in it.We started in June 2018, so we have it for a while, but I think 100,000 euros now is spent and the other half not' (Heerlen, Interview 2020).

Discussion
In this article, we took a critical view on co-production in shrinking cities.We explored the limitations that this particular urban context imposed, and analysed the instruments in spatial planning and urban governance that can be used to overcome these limitations to co-production.
The findings reveal that, although with different levels of success, practices where civil society was involved in governance or public service provision took place in each of the three cities, despite the varying strength of social ties and tradition of civic engagement.The three cases had different levels of motivation to pursue governance changes.Riga's case illustrated the deliberate attempt at governance change initiated by the civil society organization.Although the project did not result in significant power redistribution, it manifested an attempt of the NGO to strengthen their position to influence urban processes and improve access of affordable spaces for artists.Due to the lack of the municipality's involvement, co-production attempt resulted in a classic case of gentrification.In Nevers, the community centre did not attempt to challenge the municipality's planning approach, but rather focused on finding practical solutions to dealing with social problems that were ignored by municipal officials.In Heerlen, co-production of the new governance model led to a democratization of the city's governance structure, and allocated more power to civil society.It was a deliberate process initiated by the municipality to overcome their lack of capabilities to deal with degraded territories in a shrinking city.However, the residents' input was more important for the implementation phase, rather than the decision-making.For example, residents could only work on certain spatial regeneration projects and not on other spheres.Despite these differences, in all three cases we observed several communalities, which we suggest could help to overcome the limitations to co-production in shrinking cities.

The role of mediating actors
The results show that co-production efforts in three cities relied on mediating actors who initiated, moderated, and/or implemented co-production projects.In Nevers, it was the community centre; in Riga, the NGO Free Riga; and in Heerlen, the independent brokers.In shrinking cities, mediating actors not only help to overcome the lack of trust of civil society towards public authorities, but also help to allocate additional financial and human resources through a better knowledge of local assets and needs.Therefore, we consider the engagement of mediating actors as one of the tools to overcome the limitations of co-production in shrinking cities.
Our results are in line with Simon (2021), who concluded that utilizing external professional facilitators for co-production projects contributes to their successful implementation, because mediators can devote sufficient time and effort to the project, establish common rules, as well as build confidence and mutual trust with the participants.We argue that mediating actors can (1) ensure a more neutral or 'safe' space in which residents and local authorities are encouraged to develop trust, (2) stimulate creative problem-solving, and (3) guarantee that the interests of the stakeholders are heard and considered equally (Simon 2021).
However, mediating actors that are able to promote their own methodologies through strong social and human capital can interrupt community processes.This was observed in Nevers, where the top-down approach used to implement the mutual aid board resulted in a lack of ownership of the project on the part of residents.A similar issue was found in Riga, where the interests of the wider public were not considered, and mediating actors from Free Riga defined the theme of the project and the scope for participation, as well as shaping the processes to serve their own interests in the field of arts and culture in line with Turnhout et al. (2019).

The role of independent funding
Mediating actors also enable access to extra financial resources, which is very important for shrinking cities with declining budgets, and which are needed to experiment, innovate, and stimulate social change.The case of Nevers shows that financial support from the central government allowed the community centre sufficient independence to address an acute problem that was overlooked by local authorities.Therefore, external funding can be considered an instrument to reduce dependency on local government's agenda.

The role of civic empowerment and leadership in co-production projects
The three cases confirm the important role of leadership and civic empowerment for initiation and implementation of co-production projects in shrinking cities.In Heerlen and Riga, the initiators of co-production projects were in possession of strong social capital and leadership skills.The case of Nevers shows that when a co-production project was introduced in a community with a large proportion of disadvantaged people, the project lacked civic empowerment and strong leadership from beneficiaries themselves.As professional community workers compensated for this lack by taking on leadership, the project ultimately turned into a top-down approach.We conclude that empowering community groups is a prerequisite to successful co-production initiatives.Our findings are in line with Turnhout et al. (2019), who argued that the empowerment of marginal actors is generally 'a fundamental goal of co-production projects' (17).Renegotiating roles, powers, and redirecting priorities to cater to the community, is necessary to empower more vulnerable groups and involve them in decision-making.For further research, it would be beneficial to explore methods of community empowerment in shrinking cities.

The willingness of public authorities to reconsider power relations
The case in Heerlen shows that the acceptance of urban shrinkage by the municipality, as well as their ability to step back from the position of power and innovate, became central for the sustainability of co-production project.The case in Riga, meanwhile, illustrates how the collaboration between the private investor and the local NGO, with passive involvement of the local authorities resulted in gentrification and restricted the sustainable delivering of social value to the neighbourhood through co-production.The social relevance of this finding is quite significant, as the situation in which local governments favour private sector interests over community interests is not uncommon in post-socialist shrinking cities (Bosák et al. 2020;Liebmann and Fröhlich 2009;Nedučin, Krklješ, and Gajić 2019;Stryjakiewicz, Ciesiółka, and Jaroszewska 2012).
Based on these results, we suggest that the willingness of a municipality to stimulate social transformations and to take the role of 'enabler' of citizen experimentation can be a prerequisite to overcome the limitations to co-production.The shrinking context can be considered a stimulus for the municipality to experiment and open planning processes to civic society.Improving residents' access to information about public participation and adopting residents' ideas into funding schemes are tools that can help enabling successful co-participation.In this regard, further research could explore forms and processes of experimental governance in shrinking cities.

Conclusion
First, hoping to make an empirical contribution to the co-production field, we identified in the literature four limitations to co-production in diverse contexts of urban shrinkage, and tested them empirically.The first limitationthe lack of trust in public authoritieswas the main challenge for Heerlen municipality and the trigger to involve local residents in co-production of the new governance model.The second and third limitationsdiminished capacity for civic participation driven by the particular social composition and weakened social capitalwere observed in Nevers, where the lack of civic empowerment among a specific group of disadvantaged residents impeded the co-production project initiated by the community centre.On the contrary, initiators of the Riga project represented the empowered group with a strong social capitalcreative youth and artists, who were capable of initiating an urban project, and entering into negotiations with private and public actors.This fact can be associated with Riga's capital status, which is likely to affect the social composition with more young, educated and middle-class residents remaining in the city.In Heerlen, selective outmigration is not hindering co-production efforts, as the elderly comprise the main group of involved civil society actors in the Gebrookerbos project.Most of those actors have been living in Heerlen for a long time and place-attachment plays a crucial role for the tacit knowledge about the neighbourhoods.The last limitationthe dominance of the private sector in planning processes over civil societywas vividly observed in Riga, while such a constraint did not appear in the other two cases.
Second, we derived several instruments to overcome limitations to co-production.All three case studies showed the importance of engaging mediating actors in co-production processes to overcome the impaired trust of civil society towards public authorities and to allocate additional financial and human resources.Besides, the use of external funding in Nevers, allocated from sources other than local municipality, ensured independence from local government agendas and allow the project to focus on residents' needs.Leadership and civic empowerment provided a strong base for the development of co-production practices in Riga and Heerlen.Finally, the willingness of public authorities to regenerate power relations was crucial for innovative co-production of local governance in Heerlen.
The second contribution of our paper addresses the definition of co-production.We took a critical stance on the commonly used public management definition, which pursues cost-efficiency as an ultimate goal.Drawing on Mitlin (2008), Watson (2014), we adopted an approach to co-production focusing on the power dynamics in urban governance.We defined co-production in shrinking cities as the engagement of civil society in policy-making and governance, through a reallocation of decisional power, to enable participation in public service provision in such a way that benefits local residents.Empirically, the identified limitations to co-production in the contexts of shrinkage and possibilities to overcome them confirm the centrality and relevance of power reallocation processes in the success of co-production practices.
The main limitation of this study lies in the fact that the most different case study research design considers three radically different urban contexts and approaches to co-production, in terms of method, scale and goal.While this fruitfully acknowledges the empirical diversity of co-production attempts, it also limits the possibilities to effectuate a systematic comparison, in particular between similar attempts at co-production in different contexts, or conversely different approaches to co-production in a similar context, two avenues which could be logical progressions for this work.
On the other hand, the most-different case study design allows us to observe how the four limitations and four ways to overcome them were not consistently observed in all three cities, and to draw attention to the importance of local context.Shrinking cities scholars repeatedly cautioned urban practitioners and public authorities against universal planning solutions for shrinking cities (Martinez-Fernandez et al. 2015;Sousa and Pinho 2015;Wiechmann and Bontje 2015;Watson 2014).Our findings confirm that place-based instruments that consider available resources and needs of the local population should be tailored to each particular context.
Another limitation of our research design is that we cannot claim to be certain that what we observed in shrinking cities could not take place in the context of urban growth, having not conducted the comparative research with growing cities.However, we argue that in our case studies, the co-production projects were motivated by the particular needs resulted from urban shrinkage and thus are likely to be specifically relevant to this particular context.On the other hand, looking at the shrinking contexts enables to better understand the constraints and possibilities for co-production practices in the contexts of limited resources (economic, demographic, etc.).These insights could prove useful to understand limitations to co-production practices in growing areas, where actors are also often confronted to limited resources, for instance, in the context of austerity.

Notes
1. Active citizenship is understood as 'participation in activities related to political groups, associations or parties, including attending any of their meetings or signing a petition' (Eurostat Social Participation and Integration Statistics in the European Union (EU) 2021).
2. Formal volunteering refers to 'activities organised through an organisation, a formal group or a club, including unpaid work for charitable or religious organisations.Informal volunteering refers to helping other people, including family members not living in the same household (e.g.cooking for others, taking care of people in hospitals/at home, taking people for a walk, shopping, etc.), helping animals (e.g.taking care of homeless, wild animals) or other informal voluntary activities such as cleaning a beach, a forest, etc' (Eurostat Social Participation and Integration Statistics in the European Union (EU) 2021).3. Creative quarter is defined as a geographical area 'with groups of buildings adjusted to cultural and artistic industries in order to create a sense of identity and conditions facilitating and encouraging those activities' (Roodhouse 2006).4. The numbers are calculated at the municipal level.At an agglomerated level, including peripheral towns, the city counts 115,351 inhabitants (INSEE perimeter Aire d'attraction des villes, AAV). 5.At an agglomerated AAV level the percentages are respectively: 13.2%, 13.4% and 14.3%.6.One important note is that the project was interrupted twice: the first time because of the summer holidays in August, and shortly after its launch by new restrictions due to the Covid-19 pandemic.7. The International Building Exhibition (the Internationale Bauausstellung, IBA) is an experimental format of urban and regional development.It has its origins as a city planning instrument in Germany.Outside of Germany, IBAs have been held in Basel and Parkstad Limburg.

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

Table 1 .
Values of independent variables and their means per case study.
Source: own calculations based on data from Eurostat Social Participation and Integration Statistics (Eurostat Social Participation and Integration Statistics in the European Union (EU) 2021).