Features and drivers of citizen participation: Insights from participatory budgeting in three European cities

ABSTRACT

Participatory budgeting (PB) is a relatively novel approach to the allocation of funds which allows ordinary citizens to become directly involved in how local government money is spent. This study identifies and examines the features and drivers of PB that incentivize citizen participation and the co-production of public services. Our analysis takes a fresh approach by setting PB initiatives in an innovative frame combining a paradigm of ‘ideal’ types of PB and their diachronic constituent phases. The results provide insights for both scholars and policy makers on the key features and drivers of citizen participation through PB.

KEYWORDS:

• participatory budgeting
• co-production
• citizen participation
• local governments
• case-studies

Introduction

Democratization of governance comes high on the list of public governance priorities (Bingham, Nabatchi, and O’Leary 2005; Baiocchi and Ganuza 2014; Fung 2015). When deliberative models of participation (Warren 2009) are introduced, citizens become co-producers of public services (Bovaird 2007; Osborne 2010; Barbera, Sicilia, and Steccolini 2016a, 2016b; Sicilia et al. 2016). Thus, the deliberation and consensus-building associated with participatory budgeting (PB) is already seen as a central tool for fostering democracy (Bingham, Nabatchi, and O’Leary 2005). It authenticates the democratic credentials of institutions and strengthens trust between citizens and politicians (Fung 2015).

Introduced in 1989 in Porto Alegre, Brazil, PB has since been adopted in many localities both in the European Union and all over the world (Pateman 2012; Cabannes 2015; Allegretti and Herzberg 2004; Kersting et al. 2016). The involvement of citizens in public governance has been shown to counter disenchantment with political processes (Ianniello et al. 2019; Miller, Hildreth, and Stewart 2019; Strokosch and Osborne 2020). Although PB has been investigated in disciplines as diverse as urban studies, administrative science, and public policy, there has been a lack of research on the interplay between co-production opportunities and citizen participation in PB schemes (Miller, Hildreth, and Stewart 2019). Furthermore, few accounting studies have explored PB from a comparative perspective. Kersting et al. (2016) suggest that further comparative studies should focus on how contextual factors can incentivize participation in PB, especially in the design and implementation of co-production phases.

In the present study, we set out to fill some of these gaps by showing how in different ways the local governments of three Western European provincial capitals, Bologna, Stuttgart and Zaragoza, incentivized their citizens to become involved in co-production through PB. We examine the dynamics that ensued and the varying results achieved.

With particular reference to citizen participation and co-production (Osborne 2010; Barbera, Sicilia, and Steccolini 2016a, 2016b; Sicilia et al. 2016), the paper posits two research questions:

RQ1: What are the key features of PB schemes that incentivize citizen participation in the co-production of public services?

RQ2: How do drivers of PB incentivize citizen participation in the co-production of public services?

We shed light on comparative practices by combining two frameworks: the ‘ideal’ types of PB as defined by Sintomer, Herzberg, and Röcke (2008) and the constituent phases of PB delineated by Barbera, Sicilia, and Steccolini (2016b). Using data compiled from documents and semi-structured interviews from the three case studies, we show PB in action and identify the drivers of citizen participation in co-producing public services.

This research aims to contribute in several ways. First, by analysing different strands of the literature and combining different frameworks, we examine citizen participation in the context of the six ‘ideal’ types of PB (Sintomer, Herzberg, and Röcke 2008) and its constituent phases of implementation (Barbera, Sicilia, and Steccolini 2016b). Second, we focus on PB through the lens of citizen participation. Critiquing different international practices offers new stimuli for scholars to explore PB as a means of co-production (Miller, Hildreth, and Stewart 2019) and invites reflection from a comparative perspective (Kersting et al. 2016). Third, in the light of policy makers advocating the importance of citizen participation (Franklin, Ho, and Ebdon 2009; Zhang and Liao 2011; Pateman 2012), we discuss the key features and drivers under which PB leads to citizens becoming co-producers of public services. Our investigation results in several suggestions for effective PB programmes.

The paper will proceed as follows. The literature on citizen participation, co-production, and PB schemes is discussed in the next section, along with the presentation of the combined PB framework. The methodology is then outlined, after which each of the three case studies is presented in turn. The results of the comparative analysis of the key features and drivers of PB in the three cities are discussed before concluding with a summary of the findings, their limitations and avenues for further research.

Citizen participation, co-production, and participatory budgeting

Citizen participation and co-production

In public governance, the desirability of having an ever increasing number and plurality of actors involved in deciding and defining the public services to be provided is seldom, if ever, disputed (Barbera, Sicilia, and Steccolini 2016b; Strokosch and Osborne 2020). Local governments generally see citizen participation and co-production as an essential part of the structure of democracy (Ianniello et al. 2019; Miller, Hildreth, and Stewart 2019).

The concept of co-production originated in a study by Ostrom (1972) on citizens participating in the design and delivery of public services. The idea subsequently gained traction in public governance studies (Osborne 2010; Osborne, Radnor, and Nasi 2013), with Bovaird (2007) highlighting the advantages of long-lasting relationships developing between users of public services, volunteers, and community groups as co-producers with public sector organizations.

Ianniello et al. (2019, 2) further point out that ‘normative assumptions about the value of citizen participation are often taken for granted’. Without shirking the implicit caveat, this paper explores what makes participation effective. A key limitation of citizen participation is represented by the disparity between expectations and the actual opportunities to influence policy making (Michels 2012). Ianniello et al. (2019) argue that citizen participation requires adequate briefings, the support of officials, appropriate organizational arrangements including clear criteria for community representation and well-designed processes. In a similar vein, Liao and Ma (2019) emphasize that citizen participation requires careful planning, coordination, and strategic support, demanding considerable effort on the part of managers and officials. Another hindrance to citizen participation is that not all individuals wish to become involved or are even interested (Pateman 2012).

In an effort to define and delineate citizen participation, Nabatchi, Sancino, and Sicilia (2017) identify two levels of co-production: the individual level, involving public service clients and consumers (Pestoff, Osborne, and Brandsen 2006), and the broad level, where civil society organizations and other groups are the stakeholders (Bovaird 2007; Alford 2014; Loeffler and Bovaird 2018). In our study, we see co-production as a ‘heterogeneous umbrella concept’ that covers co-planning and co-design and includes both levels of participation (Nabatchi, Sancino, and Sicilia 2017, 769). Furthermore, we examine how PB promotes citizen engagement in co-production.

Participatory budgeting and its ‘ideal’ types

The first widely reported instance of PB being implemented was in Porto Alegre, Brazil, in 1989, largely thanks to the Brazilian Workers’ Party. In Porto Alegre, neighbourhoods are allocated a small portion of the city’s overall budget and decide how to spend it. The local government then accommodates and prioritizes the projects to fit with the City Council’s overarching strategy. Grassroots democracy, expressed through citizen assemblies, citizen control, and the equitable and just distribution of funds, has played an essential part in fuelling the success of Porto Alegre’s participatory budget process (Gret and Sintomer 2005; Sintomer, Herzberg, and Röcke 2008).

The end of the 1990s marked the ‘return of the caravels’, which led to the progressive adoption of PB in Europe (Allegretti and Herzberg 2004, 3). Sintomer, Herzberg, and Röcke (2008) noted four features common to PB varieties that sprang up (a) the origin of the participation process; (b) the level and type of meetings organized; (c) the type of deliberation; and (d) the position of civil society in the procedure. In addition to this, Sintomer, Herzberg, and Röcke (2008) identified six ‘ideal’ types of PB, which are essentially schemes based on a gradient of citizen involvement:

1. Porto Alegre adapted for Europe – The processes and actors in this type of PB are similar to those of the Brazilian experience. Individual citizens are extensively involved in discussions about investment programmes, the decision-making process and budget implementation.

2. Participation of organized interests – The main actors in this type are organized groups, such as civil society organizations, unions, and citizens’ associations. Participation usually centres on public policy implementation. Previous participatory traditions already exist and may lead to further consultative processes.

3. Community funds at local and city level – In this type, funds for investment do not come from the local government’s budget, and therefore the city council does not have the last word on decisions. Individuals and groups make budgetary decisions. Private companies are excluded from both funding and decision-making.

4. Public/private negotiating table – The only difference between this type and the community funds model is that private companies and international organizations can fund initiatives.

5. Proximity participation – The local government summarizes the results of discussions and randomly chosen citizens are invited to merely consult at open council meetings. This type is a development on existing structures such as neighbourhood funding groups and committees.

6. Consultation on public finances – This type is the same as proximity participation but with more emphasis on transparent municipal finances. Budgetary information is made available to the public and all citizens are allowed to express their opinion in open meetings.

The six ‘ideal’ types will provide a vehicle for the exploration of the way PB drivers work to incentivize citizen participation in the co-production of public services.

Citizen participation and co-production in participatory budgeting

One of the innovative features of PB is the way in which it involves actors from outside public administrations (Ewens and van der Voet 2019). Studies on perceptions of citizen participation in budgeting have already been carried out. For example, Franklin, Ho, and Ebdon (2009) show that elected officials favour budget preparation processes that foster closer interaction with citizens. Similarly, Liao and Ma (2019) highlight the positive role of municipal managers’ associations in fostering citizen participation. Liao and Schachter (2018) underline the importance of positive perceptions by officials of the public’s input while pointing out that sufficient professional skills to manage the process are required. Zhang and Liao (2011) analyse the influence of mayors and municipal managers in deciding whether to adopt PB, finding that mayors tend to support PB as an end in itself whereas managers support the idea only in so far as it remains affordable and citizens remain committed.

Barbera, Sicilia, and Steccolini (2016a) find that most citizens take a positive view when they are personally involved as co-planners or co-designers. But Ganuza, Nez, and Morales (2014) note that citizens’ associations play a more significant role in decision-making processes than individuals because of their capacity to mediate between different interests. Their conclusions raise the problem of how to achieve fair representation for all citizen groups. Finally, Pateman (2012) argues that the effort required to achieve consensus with citizen participation must be balanced with the benefits of developing shared ideas. Pateman (2012) concludes that the very essence of PB is found in the collaboration between public administrators and citizens and may vary according to the circumstances.

Taking their cue from Moynihan (2007), Barbera, Sicilia, and Steccolini (2016b) formulate a scheme of three broad categories for PB based on the degree of citizen participation: pseudo, partial, and full. In pseudo participation, the local government controls the entire process and makes the final decisions. Although citizens are encouraged to put forward their ideas, their involvement is essentially symbolic. Partial participation means citizens can express their ideas and preferences but their influence is limited. With full participation, citizens can have a decisive influence that is sanctioned by public officials. In all degrees of citizen participation, the essential ingredient is the citizens’ willingness to invest their time and competencies in preparing proposals.

Co-production in PB entails co-planning and co-designing (Barbera, Sicilia, and Steccolini 2016a, 2016b). An important feature of co-planning is a collaboration between public sector organizations and individuals and citizen groups to decide which public services should be funded (Bovaird and Downe 2008). Co-design involves defining how the selected plans will be delivered (Ostrom 1996; Osborne 2010). Thus, PB can be an institutional link between citizens and the centres of decision-making in local governments (Baiocchi and Ganuza 2014).

Citizen participation and co-production in different types of participatory budgeting

Our integration of Sintomer, Herzberg, and Röcke’s (2008) framework incorporates Barbera, Sicilia, and Steccolini’s (2016b) three degrees of participation (pseudo, partial, full). Partial participation schemes proposed by Sintomer, Herzberg, and Röcke (2008) include (1) Porto Alegre adapted for Europe and (2) participation of organized interests. There are also two full participation schemes: (3) community funds at the local and city levels and (4) public/private negotiating committees. Finally, (5) proximity participation and (6) consultation on public finances are recognized as pseudo participation schemes. In our analysis, we proceed with this combined framework, which we call PB schemes. Using our enlarged framework, we investigate:

RQ1: What are the key features of PB schemes that incentivize citizen participation in the co-production of public services?

The effectiveness of using PB as a means for engaging citizens has also been discussed at length. Scholars typically measure effectiveness in terms of the number of citizens involved (Ebdon and Franklin 2004), the representativeness of those involved (Franklin, Ho, and Ebdon 2009; Hong 2015), and their actual impact on budgetary decisions (Ebdon and Franklin 2004; Watson, Juster, and Johnson 1991). For Hong (2015) the level of efficiency is the percentage of citizens’ proposals that are adopted and carried out. For citizen participation to be effective it needs to be rooted in dialogue, as ‘two-way communication is important for conveying what is desired and what is possible given resource constraints’ (Franklin, Ho, and Ebdon 2009, 55). Drawing on cases of three European cities, we use the PB schemes as a vehicle to explore the question:

RQ2: How do drivers of PB incentivize citizen participation in the co-production of public services?

Research methods and context

The research questions are qualitatively addressed by using comparative case studies which are of the utmost value in stimulating enquiry, developing theories, providing examples, and problem-solving (Cooper and Morgan 2008). According to Starbuck (1993), to understand the specific features of a process as opposed to general traits that could be found in any organization, one must seek practical perspectives by observing systems as they develop. To date, comparative case studies on PB have been conducted with both cities in the same country (Wu and Wang 2012) and in different countries (Ganuza, Nez, and Morales 2014). Previous research had been only loosely related to citizen participation and co-production, leaving scope to explore key features and drivers of PB and draw conclusions regarding how best to foster citizen participation in the co-production of public services.

Case selection and context

Theoretical sampling was applied in the choice of the case studies (Eisenhardt 1989). The countries were selected following the ‘Most Similar Systems Design’ (Przeworski and Teune 1970), which aims to keep as many significant contextual factors as constant as possible to assess the impact of the key causal factors differing between the cases. Germany, Italy, and Spain were considered suitable as they have relevant similarities. (1) The three countries are members of the European Union and have the same administrative culture based on the Rechtsstaat (rule of law) (Pollitt and Bouckaert 2017). Rechtsstaat states are based on the supremacy of a national constitution and guarantee the safety and constitutional rights of their citizens. (2) In contrast to smaller countries, Germany, Italy, and Spain are ‘too big to fail’ (Badell et al. 2019), as their size would make a full bailout practically impossible. This has implications regarding financial stability, which can affect the allocation of PB funds. (3) The selected countries show a similar level of confidence in the civil service (World Value Survey 2020), which represents a key element in explaining citizen satisfaction and trust (Pollitt and Bouckaert 2017).

Italy and Spain are both characterized by a pronounced separation between State and society, clientelism, party patronage, centralized government with relatively weak local governments (Pollitt and Bouckaert 2017). In both countries, PB can be linked to national public management reforms that attempt to set up processes with clear rules for cooperative local governments with active citizens (Sintomer, Herzberg, and Röcke 2008). In Italy, the reforms of Legislative Decree No. 267/2000 set out several types of citizen participation, and different municipalities responded by appointing city councillors with responsibility for PB. In Spain, local governments have financial autonomy, and most of them make use of representative bodies to encourage citizen participation in political decisions. Germany, on the other hand, is a country with greater cohesion between State and society, a decentralized federal structure and strong local government (Pollitt and Bouckaert 2017).

As PB is a local experience, we focused our attention on local government, searching for large entities on the understanding that organizational complexity can favour participatory innovation. Furthermore, size has an indirect effect on innovation because it fosters functional differentiation and, in turn, organizational complexity (Ewens and van der Voet 2019). We reviewed PB experiences in each country and double-checked for the availability of documents online. We then selected one case per country, applying criteria of size in terms of local governments, population, regional importance and economic production. Bologna, Stuttgart, and Zaragoza met our criteria: they are large municipalities (see Table 1), regional capitals of Emilia-Romagna, Baden-Württemberg, and Aragon, and are key players in the economic development of their respective regions with GDP per capita higher than the national average. Plus, they all have a similar productivity profile (Eurostat 2020), with about 64–72% of regional GDP in the tertiary sector and about 25–28% in the secondary sector.

Table 1. Interviewees with their different roles in the participatory budgeting process.

	Bologna (B)	Stuttgart (S)	Zaragoza (Z)
Local government actors	Government official in charge of the budget (Interviewee B1)	Government official and Participation Advisory Board member (S1)	Government official in charge of the budget (Z1)
	Government official in charge of PB process (B2)*	Government technician in charge of PB process (S2)
		Governmental head of the public relations department (S3)
External actors with a key role in implementing participatory budgeting	Head of the foundation in charge of PB, University of Bologna (financed by the City Council) (B3)	Member of the voluntary working group (S4)*	Chair of the Citizen Participation Laboratory, University of Zaragoza (financed by the City Council) (Z2)*
	Member of the foundation framing and monitoring the PB (B2)*	Member of the voluntary working group (S5)*
Citizens and their representatives	Expert citizen (B4)	Member of the voluntary working group (S4)*	Expert in citizen participation (Z2)*
	Representative of a neighbourhood association (B5)	Member of the voluntary working group (S5)*	Representative of a neighbourhood association (Z3)
	Chair of the district workshop in charge of the administrative process (B6)	Representative of a citizen association (S6)	Secretary of a neighbourhood association (Z4)

Note. *: Interviewees acting in the process with different roles

When we interpreted our data, we focused on citizen participation in co-planning and co-designing (Osborne 2010; Barbera, Sicilia, and Steccolini 2016a, 2016b). We assessed the extent to which the case experiences fit the PB ‘ideal’ types (Sintomer, Herzberg, and Röcke 2008) and through this analysis empirically discussed how PB incentivizes citizens to participate in co-planning and co-designing public services.

Data collection

We collected data from multiple sources in order to analyse PB in the three chosen cities. This approach also helped to triangulate the data. Documentary evidence was sourced from websites, newspaper articles, and municipal archives. To reflect on the roles, interactions, and perceptions of various actors in the PB process, we conducted individual semi-structured interviews. In each of the three municipalities, interviews were held with politicians, officials, and citizens between April 2018 and March 2020. Each interview lasted around 90 minutes with at least one researcher present. The sessions were recorded and key statements were transcribed. A list of interviewees is provided in Table 1.

Each phase of the PB process was examined for each municipality. We paid particular attention to citizen participation through co-production including co-planning and co-design (Barbera, Sicilia, and Steccolini 2016b). The results of our analysis are presented in the next section.

Citizen participation in participatory budgeting: analysis of three cities

The case study for each city follows, after which we identify and discuss the key features of citizen participation.

The Bologna case

Bologna is the seventh-largest city in Italy with a population of 389,000. It has a long and unique tradition of left-wing local government, with several extra-electoral participation structures, including citizen assemblies, public meetings, and focus groups, which make for consistent participation patterns (Stewart 2007). Grassroots democracy involving both citizens and politicians has long been a part of the Bologna experience. In 2017, the Municipality of Bologna financed an initial PB project for each of its six districts at the cost of €1 million.

The inception. In Bologna PB may be seen as the natural evolution of a process already underway and initiated by citizens and associations (as well as politicians) as part of an active search for opportunities to participate. Steps towards citizen participation in the municipal budget started in 2010 and led to the creation of a Foundation for Urban Imagination in 2013. This group runs the district workshops where citizens discuss and select areas for PB intervention. The creation of a dedicated institutional body represents a structural change that has improved the entire process (Baiocchi and Ganuza 2014). As several interviewees remarked, district workshops are the essential channels of citizen participation, ensuring constructive dialogue between the citizens, associations, and government representatives:

To get people and the administration to really work together, Bologna needed a collaboration tool to facilitate dialogue between the municipality and the citizens. That is why we set up the district workshops, one in each district, to discuss eleven already-prioritized areas of intervention identified by the citizens, among which six were selected for PB. (Interviewee B2)

The participation process. Groups of citizens engaged in planning and implementing the projects with the municipality. The Municipal Council, together with the Conference of the District Council Presidents (elected by the citizens of the district), settled on the rules and procedures for the budgeting process. The conference monitored progress, while each district council conducted the workshops, following a direct participation approach (Cabannes 2004).

Implementing the budgeting process was divided into three phases. The first phase was co-planning (through the workshops) to identify critical issues, opportunities, and objectives and to reach consensus on future actions among the citizen participants, their associations, district teams, and governance units of the municipality, including the supervising foundation.

[The phase] followed the logic that the leadership and guidance of the process over the years will have to move towards even greater autonomy for citizens. (B6)

This phase considered the first step towards effective participation (Ebdon and Franklin 2004), in Bologna relied on workshops, with 331 citizens and 134 associations exchanging knowledge, analysing resources, assessing local development projects and so on. Researchers have shown that including associations in this process can make a significant difference in city participation (Liao and Ma 2019).

In the second phase, proposals were gathered, and co-design began. New spaces for collaborative activities using PB resources were identified, and efforts were made to reach out to young people and migrants. Again, co-design took place in the workshops, but also in informal meetings, with 664 citizens involved in the process. In accordance with Ebdon and Franklin’s (2004) definition of effective PB, deliberately involving as many citizens as possible increased the effectiveness of the entire process. Moreover, the level of engagement was perceived by both the leaders of the project and the public officials as a significant step towards inclusiveness and democratization.

In the third phase, co-design took place in meetings with 225 citizens working on PB proposals alongside technicians and facilitators. These collaborations resulted in 84 proposals. Of these, 36 concerned interventions already financed from other regional resources and for that reason were excluded from the PB initiative, leaving 48 proposals.

Citizen participation affected the allocation of funds in several ways. The dialogue between a broad cross-section of parties surrounding co-planning and co-designing was an integral part of formulating priorities, policies, and projects.

The deliberation process. The deliberation process had a high level of citizen involvement with representatives of both citizens and LG formulating the investment plan. Citizens presented proposals as part of public discussions or submitted them to the website, which were then often further tweaked at the district workshops. Proposals with a go-ahead by the municipality in terms of technical and economic feasibility, sustainability and timing were put to an online vote in an attempt to reach as many citizens as possible. A communication campaign was activated in each district to promote the deliberation process:

A communication campaign was launched: ‘If you have lost faith in the system, get involved. Your district begins with you.’ And we took great pains to reach every citizen. We sent people to each district with tablets to get people walking along the street to vote. (B3)

The remaining proposals for projects (48), were reduced to 27 after technical analysis. No further deliberation was required at this stage.

The final decision. Citizens took the final decision. All citizens aged over 16 were eligible to vote, including migrants, residents, and non-residents living, working, or studying in Bologna. Notably, citizens could vote on any proposal in any district of Bologna, not just in their own area. The final vote count was 14,584 people spanning all age groups, although turnout among the young was lower. Sixty-one percent of voting citizens voted for projects in their district of residence. About one-third of the voting citizens already had an account on the municipal website, reflecting a pre-existing connection to the municipality. The full budget for each district, about €150,000, was invested in the proposal with the highest number of votes. The six ‘elected’ projects commenced at the beginning of 2018. Each funded project was given a page on the community website to report progress to citizens. Despite the relatively small amount of funds available to each district, the key actors perceived the budgeting process as a successful exercise in citizen participation:

The benefit of the entire process is reflected in the 3% of the population that decided to participate. Citizens are ready to collaborate in the decision-making process of the municipality. (B5)

Following citizen participation in the 2017 budget process, the City Council decided to use citizen participation again for the 2018 budget. In particular, they sought to target previously under-represented communities such as young people and immigrants. They also increased the amount of funds available for investment. The success of PB in Bologna is reflected in the words of an interviewee:

It was surprising to see so much participation in the participatory budget, given the limited resources involved. In response to the citizens’ enthusiasm, the council decided to increase the amount devoted to the initiative. (B1)

The Stuttgart case

Stuttgart is the sixth-largest city in Germany with a population of 635,000. Since 2013 the local government has been headed by the Green Party. A major civil uproar against a large railway construction project called ‘Stuttgart 21ʹ in the years 2008–2010 gave rise to a demand for citizen participation at all levels, leading to a broad citizen participation programme. Independently of that programme, the first PB effort was implemented in 2011, and it has subsequently been held in two-year cycles. Thus in Stuttgart PB, originally a response to a troubled citizenry, was transformed into a mainstream process.

The inception. It was initiated by both citizens and politicians and based on a broad political consensus. The Treasurer’s Office developed a proposal for a ‘PB learning’ process that was approved by the City Council in March 2011. It allows for adjustments to each PB cycle to increase representativeness and transparency.

From the outset, … there has been a PB Task Force with the mayor, representatives of each political faction, representatives of the adult education centre, and a working group of volunteers, who jointly evaluate the previous PB process, and make suggestions for improvements. (S4)

Our analysis focuses on the fifth PB cycle which began in January 2019.

The participation process. Citizens were able to submit proposals, comment on them and vote. A group of volunteers engaged in information events and training seminars assessed the previous PB cycle and suggested improvements in the co-design of the next one.

The cycle began with a marketing campaign and a series of information sessions about upcoming activities. These sessions included district meetings and three training seminars organized by the working group, designed to attract new supporters and participants. In 2019, each seminar targeted a different previously under-represented group: the elderly (in conjunction with the City Seniors’ Council), migrants (with the Forum of Cultures), and youth (with the City Youth Council).

In the proposal phase, all Stuttgart residents (irrespective of age or nationality) were given three weeks to prepare proposals. There were no restrictions on the type of proposal (e.g. size, extent, purpose) nor on the cost (proposals were not required to include cost estimates). An external agency parsed all the proposals and bundled together submissions on similar topics to reduce the number of voting alternatives. This action was decoupled from the authority of the city of Stuttgart, which also did not play a role in any feasibility checks at this time.

Although two-way communication is considered an effective driver of PB (Franklin, Ho, and Ebdon 2009), in Stuttgart communication during both phases was largely one-way, with city councillors simply informing the citizens of developments and decisions. Nonetheless, the process allowed for wide representativeness and inclusiveness: the number of proposals submitted dramatically increased since PB was launched, from 1,745 in 2011 to 3,753 in 2019.

The deliberation process. During the voting phase, residents of any age could comment on and/or vote in favour of or against each proposal. A total of 40,620 residents cast 1.4 million votes regarding 3,753 proposals. In 2019, district workshops were added to the PB process for the first time, and took place in 17 of the 23 districts. Representatives of the city administration and district councils attended the workshops. Theoretically, this opportunity for two-way communication should have led to greater effectiveness (Franklin, Ho, and Ebdon 2009). However, the workshops failed in this regard, as the citizens largely ignored the meeting.

Citizens may talk to each other, trying to prioritize proposals and to bring out the main points … But the events were rather poorly received … At times, there were more city officials than citizens. We will have to see whether we will proceed with the workshops. (S2)

An easy online registration process, with only random controls and no barriers, was also designed to encourage citizen participation. Opinions differed between government and citizens:

We found the easy access to the platform somewhat problematic. But this also has to do with the non-bindingness of the participatory budgeting process. (S1)

Some citizens were about to participate through the online platform. They stopped because they did not want to give personal data. (S6)

On the assumption that citizens are usually more reluctant to provide their personal details in writing, some proposers started to register residents by collecting signatures in the streets. These signature lists were then submitted to the city administration, which manually checked whether each signatory was eligible to vote as a resident of Stuttgart.

Overall, the ease of access to voting, improved over several budgeting cycles, was seen as a contributing factor to the representativeness of the initiative, which is another key feature of successful citizen participation in PB (Franklin, Ho, and Ebdon 2009).

The proportion of women was pleasingly high, i.e. more than 50% of the people who voted – that is not always the case in political processes. … For us, it showed that PB is a good way to get citizens involved. (S5)

The final decision. Citizens did not take part in the decision-making process, which hindered their ability to co-design approved plans. However, proposals voted on indirectly affect political decisions, as politicians could not simply ignore the choices voted on by citizens during the PB process.

No political group can afford to make decisions without taking into account the participatory budget. So, there is pressure, not only from the PB working group of volunteers, which is politically neutral, but also from the citizens themselves [that the proposals should be implemented]. (S4)

It is worth noting that the voting system itself was open to being skewed, in that winning proposals could be the result of intensive campaigning rather than usefulness. The 100 top-voted proposals and the two with the highest number of votes per district were published on the website. Also published were the positive and negative results of any technical assessments undertaken and comments by citizens as well as the city district advisory boards. Proposals beyond the scope of the budget were not eliminated, but instead directly forwarded with a ‘yellow card’ to the ideas and complaints system of Stuttgart as part of the afore-mentioned broad citizen participation programme. This procedure extended the role of citizens as co-planners beyond the PB process.

The remaining steps of the decision process were not transparent, and there was no further interaction with citizens. Councillors could choose whether or not to include or name proposals in their contributions to the budget debate at the City Council. As in the past, there was no formal obligation to approve any of the proposals or to account for why proposals were refused or ignored. Even information about the implementation of approved projects was hard to come by and usually only disseminated as exemplars when promoting the next PB cycle.

Despite the limitations (e.g. one-way communication, co-planning only, opaque decision-making), all actors perceived the PB project as effective due to the high number and quality of the proposals, the low barriers for proposing and voting, the increased representativeness of the participants because of targeting, and the co-designing role of volunteers.

The Zaragoza case

Zaragoza is the fifth-largest city in Spain with a population of 660,000. Over the last 15 years, both central and local government have attempted to increase citizen participation in the city’s governance. In a series of reforms to address the issue, the municipality was divided into 14 districts, each with a district board and a neighbourhood committee to promote and expand citizen involvement.

The inception. The tentative steps towards transformation from traditional budgeting sprang from a new style of governance driven by politicians, but not broadly supported by all parties. The left-wing political party implemented PB in 2017 as part of a suite of transparency and participation policies; the right-wing did not agree. Thus, there was no broad-based political support at the launch of PB, which made the institutionalization of participatory practice difficult (Kersting et al. 2016).

The participation process. The main objective was to involve citizens in co-planning municipal policies and to help prioritize citizen needs. As with other local government experiences, organizational changes were made to improve the process (Baiocchi and Ganuza 2014) – for example, the creation of a municipal unit and a local political leader responsible for managing the programme, plus a technician and officials to supervise the process.

The PB system followed a direct participation approach that involved participation by neighbourhood committees and citizens in co-planning (Cabannes 2004). Design was mainly carried out by the local government. The district boards were in charge of promoting, coordinating, and managing the process, and for guaranteeing the rights of those eligible to participate. A participatory committee in each district organized open forums that were held regularly during each phase of the process. These forums dealt with general information, self-diagnosis, deliberation on proposals, prioritization, and ensured two-way communication, the latter being a key driver of PB (Franklin, Ho, and Ebdon 2009).

The deliberation process. Citizens could submit proposals and vote on them, attesting to their role as co-planners. In 2017, 1,484 proposals were received, 56% of which were prepared by citizens’ associations and the remainder by individuals. Rivalry between individuals and associations is common (Ganuza, Nez, and Morales 2014) and carries the risk of the latter dominating procedures. Of the proposals submitted, 23% were rejected because they exceeded the funding limits or suggested projects beyond the scope of the municipality’s purview, leaving 1,145 proposals for initial evaluation by the participatory committee. Of these, 392 proposals were prioritized for review by the technical team and 211 were deemed viable. The high percentage of rejected proposals raises questions about the effectiveness of the co-planning process. Providing citizens with more information and support would be likely to increase the proportion of feasible submissions:

When you make a proposal, you do not have enough information about viability or whether the proposal makes sense in the municipal context. (Z4)

This weakness in co-planning supports Barbera, Sicilia, and Steccolini (2016b) argument for more interaction between citizens and the government to make the process more effective.

The final decision. A total of 6,132 citizens voted, which equated to 1.07% of the total number of eligible voters. They selected 120 out of the 211 projects deemed viable, including some with very few votes.

Reactions to the PB initiative have been varied. Political leaders and the council technicians considered the process to have been very successful.

Everyone who wanted to participate was able to do so. The final voting procedure led to the final decision. (Z1)

The interviewees also generally confirmed that the voting process was a decisive break from traditional decision-making, where there was little or no place for the opinions of citizens. However, some interviewees were cautious in labelling the exercise a success despite the overall positive perception of the co-planning concept:

Citizens were given the opportunity to participate in decisions about the allocation of a part of the budget … . but only 1% participated. (Z2)

The participation in the decision about investments was positive, but my concern is about the level of participation. Only 1.07 % of the citizens participated in the process. Some people were unable to participate either in person or online. (Z3)

Although citizens voted on proposals, they were not subsequently involved in the development of the projects. According to Barbera, Sicilia, and Steccolini (2016b), this leads to low levels of participation in co-designing activities. As it is part of the local government’s responsibilities to develop the approved proposals, one of the main problems that emerged was a lack of two-way communication during this phase.

The administration continues to be responsible for providing the services and developing the projects selected. (Z2)

There is little interaction between citizens and administration, as there is no further involvement of citizens in the development of the project. (Z3)

In the first year of the initiative (2017), 92.4% of the projects were completed, but the unfinished projects raised questions about the viability of annual budgeting. It was therefore decided to move to a biennial schedule, so PB and implementation would take place over a two-year period. In Zaragoza, the 2019 changeover of political parties in local government saw the PB experience abandoned. This happened with many other Spanish local governments. Political division makes institutionalizing participatory practice difficult (Kersting et al. 2016). The new local government administration is still considering whether to reintroduce PB in 2021 in a way that assures greater citizen involvement.

These findings are summarized in Figure 1 and discussed in the following section.

Figure 1. PB in Bologna, Stuttgart and Zaragoza.

Source: Authors’elaboration on Sintomer, Herzberg, and Röcke (2008) and Barbera, Sicilia, and Steccolini (2016b)

Findings: key features and drivers of participatory budgeting

From the case analyses, it is possible to identify the key features of PB in co-planning and co-design processes (RQ1). First, the introduction of PB appears to be partly sparked by a political will to increase citizen participation. The inception phase of Bologna’s experience falls into Sintomer, Herzberg, and Röcke (2008) ‘ideal’ type of ‘participation of organized interests’, whereas Stuttgart and Zaragoza are more aligned with the ‘Porto Alegre adapted for Europe’ type.

A second feature is the creation of a separate body to take charge of and foster co-production. In Bologna district workshops were used to coordinate co-planning and co-design. In Zaragoza a participatory committee was responsible for the coordination of the proposals. In Stuttgart the citizen volunteers took part in the PB Task Force. Stuttgart’s PB exercise aims to generate the maximum number of proposals, regardless of feasibility, even though subsequently citizens are excluded from full co-planning and all co-design. While the limited co-planning reduces the opportunities for citizens to co-produce, we consider that ‘consultation or information is better than no consultation or no information’ (Pateman 2012, 14). Thus, the Stuttgart experience can be seen as a move towards citizen participation in the fostering of co-planning opportunities, but not full co-production.

The third key feature is the flow of information, both for soliciting proposals and selecting the ones to be put to a vote. In all three cities, the initiative generated enough interest and knowledge to solicit a significant number of proposals. Multiple information channels were established to promote and retain wider citizen participation.

The fourth feature was represented in the organizational arrangements for the presentation of proposals. Bologna and Stuttgart fostered wide community representation; Zaragoza did not. Through the lens of Sintomer, Herzberg, and Röcke (2008) ‘ideal’ types, Bologna’s experience reflected the ‘community funds at local and city level’ type, where groups of citizens engage as partners with the government in co-planning and co-designing projects. Stuttgart’s experience was more aligned with the ‘participation of organized interests’, with citizens co-planning future services but not involved in co-design. Zaragoza corresponded to the ‘Porto Alegre adapted for Europe’ since citizen involvement in decision-making was limited to defining priorities and choosing representatives. The experience confirmed previous evidence on the role of transparent communication and community-wide representation (Ianniello et al. 2019).

The fifth feature is seen in the deliberation procedure. In all three cities, all residents were allowed to take part in the deliberation process. In Stuttgart there was no age limit, while in Bologna and Zaragoza an age limit was set. Bologna and Zaragoza had similar deliberation procedures, where representatives of assemblies and local governments formulated the investment plan, as in the two ‘ideal’ types of ‘Porto Alegre adopted for Europe’ and ‘community funds at local and city levels’. Stuttgart set up district workshops in its 2019 budget cycle, although without broad acceptance by citizens, thus resembling ‘consultation on public finances’. However, participation has increased since the first run due to the efforts of volunteers. Access to information has also become easier. According to Barbera, Sicilia, and Steccolini (2016a), all of these approaches foster inclusiveness.

The last feature relates to the final decision. In Bologna, the final decision was in the hands of citizens, incentivizing involvement in co-planning and co-design, thus placing Bologna in the ‘ideal’ type of ‘community funds at local and city levels. However, citizens in Stuttgart and Zaragoza were excluded from this last step. The final decisions in Zaragoza were made by public officials and chosen interest groups, as in the ‘Porto Alegre adapted for Europe’ type. In Stuttgart, the City Council made the final decision alone (and in a less than transparent manner), reflecting the ‘consultation on public finances’ type.

From the analysis of the PB cases, we deduce that citizen participation in co-production is ‘full’ in Bologna, ‘partial’ in Zaragoza, and ‘pseudo’ in Stuttgart (Barbera, Sicilia, and Steccolini 2016b).

The key features identified in the three cities are summarized in Table 2.

Table 2. Key features of the participatory budgeting processes for each case study.

	Bologna Italy	Stuttgart Germany	Zaragoza Spain
GENERAL
Population at time of PB	389,000	635,000	660,000
Number of districts	6	23	14
Political party	Left-wing	Green	Left-wing
Participatory budgeting frequency	Annual	Biennial (like the budget)	Annual
PB cycle	1	5	1
PB proportion of total budget (%)	0.2%	Flexible	0.7%
Funds allocated to PB per inhabitant (€)	2.57€	Flexible	7.57€
Eligibility to vote	All residents over 16 (plus not resident workers and volunteers)	All residents of all ages	All residents over 14
THE INCEPTION
Inception key players	Politicians, association, and citizens	Citizens	Politicians
Project development leaders	Municipal technicians and external public foundation	Municipal technicians and external working group	Municipal politicians and municipal technicians
Key implementers	District councils	Ex-head of the City Treasurer’s Office; working group	District boards; district participatory committees
THE PARTICIPATION PROCESS
Key players	Citizens and associations	Citizens and external working group	Citizens and participatory committees
Tools	District workshop, informal meetings	District meetings, seminars	District forums
Facilitators	Municipal technicians, to a great extent	Volunteers	Municipal technicians, to a low extent
Number of proposals	84	3,753	1,484
THE DELIBERATION PROCESS
Presentation of proposals	In person and online	In person and online	In person and online
Feedback (from other citizens)	In person and online	In person and online	In person and online
Kind of assessment	Technical and economic feasibility	None (only a merger of similar proposals)	Technical and economic feasibility
Proposals remaining	27	2,901	211
THE FINAL DECISION
Proposals approved by vote	6	Unclear (but 130 proposals submitted to the Municipal Council)	120
Implementation of the most voted proposals	Compulsory for the Municipal Council (for all the winning proposals)	Not compulsory for the Municipal Council, irrespective of the vote	Compulsory for the Municipal Council (for all the winning proposals)

A number of important findings emerge for theory and practice. With respect to RQ1 on the key features of PB schemes that can incentivize citizen participation, we found results that bring previous evidence into question. The Bologna PB scheme of ‘full’ participation contained all the elements proposed by scholars to ‘guarantee’ effective PB (Barbera, Sicilia, and Steccolini 2016a; Ianniello et al. 2019), partly thanks to the city’s unique long-standing culture of citizen participation. Yet engagement rates, even if they are in line with the standard experience of between 2% and 7% of the population, as reported by Cabannes (2004), are strikingly low. When only a small group of citizens acts as co-planners and co-designers, the impact of PB as grassroots democracy needs to be put in perspective. More investigation of the Bologna case may clarify what could be done to involve a larger proportion of the population. For instance, a hybrid system could be adopted encouraging roughly sketched ideas as well as more fully elaborated proposals. Both kinds of proposals could be shortlisted for voting, thus involving more citizens in co-design.

A ‘partial’ participation scheme of PB was identified in Zaragoza and Stuttgart. In Zaragoza, the initiative was innovative and top-down, aimed at legitimizing government. However, the municipality did not offer enough support to citizen participation. Furthermore, information and participation channels were lacking or defective.

One seeming anomaly in our classification would appear to be Stuttgart’s participation rate of 6.4%, which was by far the highest of all three cases. We attribute this to several factors. First, there were no restrictions on the size or practicality of proposals, inviting any number of speculative proposals given the low bar for entry. Second, the PB Task Force has a prominent role in co-designing the next PB cycle – a strategy aimed at incentivizing the participation of all citizen groups. Finally, in Germany, by law, the City Council has the final say on budgetary decisions. Nonetheless, Stuttgart’s residents are willing to take advantage of the opportunity to express their ideas and priorities, thanks to the low barriers for contributing. The situation is quite the opposite in Bologna and Zaragoza: even though citizens there have higher expectations regarding the potential impact of their ideas and priorities, participation rates are low.

With regard to RQ2 concerning the drivers of PB to incentivize citizen participation in the co-production of public services, our findings reveal that some key drivers play a significant role. An important driver is the long-standing tradition of democratic participation, which in Bologna prepared the way for the commitment of citizens and their associations in the preparation of proposals. The guidance ensured by a specialized body designated to elicit and coordinate proposals (as well as assess their feasibility) may also be considered a driver of co-planning and co-designing, confirming the findings of Baiocchi and Ganuza (2014). For instance, the working group of volunteers and PB Task Force in Stuttgart play a fundamental role in co-designing improvements for successive budgeting cycles. Regarding the selection of proposals to put to a vote, in Bologna officials progressively filtered proposals by helping citizens to hone their priorities with a two-way communication system. In Zaragoza, to a lesser extent, officials provided similar input. However, in Zaragoza the role for the participatory committees was limited and there was a lack of coordination between the administration and citizens. This confirms that PB requires specialized guidance (e.g. by officials) (Liao and Ma 2019). Resources were spread too thinly over numerous projects, and the capacity for co-design was limited. In Stuttgart, the officials did not provide any input into the proposals, and the opportunity for a two-way dialogue – as an effective driver of PB (Franklin, Ho, and Ebdon 2009) – was largely neglected.

Finally, inclusiveness is another key driver of citizen participation, although not necessarily linked to co-production. It makes a direct address to all citizens. In the case of Stuttgart, the concept of a ‘PB learning process’ enabled citizens to participate with varying degrees of involvement, bringing to bear personal resources and skills in designing proposals. Viewed from this angle, the Stuttgart PB was, in fact, effective. Nonetheless, to lead to citizen co-production, inclusiveness has to be extended to the deliberation process and the final decision (Barbera, Sicilia, and Steccolini 2016a). Citizen involvement in the final decision is a strong driver for citizen co-production, and encouraging it can foster democratic participation and create long-lasting collaboration between citizens and governments (Bovaird 2007). Inclusiveness can also be fostered by citizens’ associations, as in the cases of Bologna and Zaragoza.

The lack of drivers can limit or interrupt the PB experience. Such was the case of Zaragoza, where a limited tradition of citizen participation, lack of specialized guidance, coupled with limited political will, resulted in a one-shot experience.

Conclusions

This paper contributes to the literature on PB by examining schemes and key features of PB, and by identifying drivers that can incentivize citizen participation in co-planning and co-designing public services. Our findings may provide scholars, technicians, and policy makers with valuable insights into the patterns inherent in PB. The main patterns of PB revealed in three local governments provide evidence that the PB ‘ideal’ types proposed by Sintomer, Herzberg, and Röcke (2008), can indeed only partially coincide with concrete experience. Only the case of Zaragoza came close to resembling one of the ‘ideal’ types (‘Porto Alegre adapted for Europe’). Our results leave space for further speculation on how PB programmes can challenge predefined schemes, rather than requiring adaptations for closer alignment to them. We also find that citizen participation can be decoupled from co-production, thus bringing into question aspects of the presumed role of PB.

However, certain drivers continue to be considered essential for the effectiveness of the process. These include collaboration between citizens and public officials – for instance at design level – and the adoption of enhanced information and participation modes. Furthermore, our findings show that when a municipality fails to take part in a genuine two-way dialogue, co-planning activities dissolve, and further efforts will be needed to involve citizens as co-producers for the PB process to realize its full potential.

The research offers signposts for scholars, politicians, and technicians in local governments, as it identifies key drivers of citizen participation to incentivize co-production. Our findings may also be useful to citizens who wish to participate in shared decision-making processes such as PB. Introducing PB and increasing participation in democratic practices requires commitment from local government with skilled actors providing appropriate settings. Equally important is providing citizens with guidance on how best to exploit PB for the benefit of their communities. However, citizens’ willingness to devote their time and competences to designing the proposals and prioritizing the ones they prefer is also essential and cannot be taken for granted. Without this vital input PB would be a mere gimmick.

While our research has several positive theoretical and practical implications, it is not without limitations. First, our findings are derived from comparing and contrasting cases located in three different countries, with different backgrounds of citizen participation in governance. An in-depth analysis of the PB experience has been carried out for these localities. However, any kind of generalization beyond the investigated contexts must be considered with caution. Second, in both Bologna and Zaragoza only the first-year experiences of PB were examined, whereas Stuttgart was in its fifth run. Future research would be able to explain better the drivers of PB after inception and/or over time. Furthermore, a comparison with other countries and cultures would provide new insights. A deeper investigation into the reluctance of citizens to take part in PB might also illuminate as yet unexplored conditions that hinder increased participation and inclusiveness.

Acknowledgments

The authors are grateful to Paul Gidwani for his invaluable support in editing.

Disclosure statement

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

Additional information

Funding

Th authors acknowledge the Department of Management, University of Bologna, Italy for financial support under MIUR Excellence Department initiative

Notes on contributors

Francesca Manes-Rossi

Francesca Manes-Rossi is an Associate Professor of Accounting at the University of Naples Federico II, Italy, where she teaches and conducts research on accounting and auditing both in the private and public sectors. She has published articles and book with international editors and coordinates research activities in public sector accounting. She is co-chair of the XII- Permanent Study Group of the European Group of Public Administration (EGPA).

Isabel Brusca

Isabel Brusca is a Professor of Accounting in the Department of Accounting and Finance at the University of Zaragoza (Spain). Her research and professional interest are focused on financial reporting in the public sector, performance management in non-profit organizations and non financial reporting. She has participated in numerous research projects in the field and is the author of several books and papers in prestigious journals. She is co-chair of the XII- Permanent Study Group of the European Group of Public Administration (EGPA).

Rebecca Levy Orelli

Rebecca Levy Orelli is an Associate Professor of Accounting at the Alma Mater Studiorum University of Bologna. Her research interests lie in the field of accounting in the public sector and accounting as social and institutional practice. She focuses on social and environmental performance management and measurement. She extensively published her work in many international journals and books. She leads the Accountability section at the Institute for the Public Value CerVAP at the University of Ferrara.

Peter C. Lorson

Peter C. Lorson is a Full Professor holding the Chair of Financial Accounting, Auditing and Management Control and Executive Director of the Center for Accounting and Auditing at the University of Rostock, Germany. He was coordinator of the EU-funded Erasmus+ project DiEPSAm and is now heading the EU-funded Interreg Baltic Sea project ‘EmPaci’. Currently, his preferred research fields are the comparison and convergence of private and public Sector Financial and Non-Financial (Sustainability or Integrated) Accounting/Reporting as well participatory budgeting approaches.

Ellen Haustein

Ellen Haustein is a post-doctoral researcher and lecturer at the Chair of Accounting, Auditing and Management Control at the University of Rostock, Germany. She obtained her doctoral degree in management accounting at the University of the West of England in Bristol. She coordinated the EU-funded project ‘Developing and implementing European Public Sector Accounting modules’ (DiEPSAm) and is coordinator of the EU-project ‘Empowering Participatory Budgeting in the Baltic Sea Region’ (EmPaci).