Introduction

The concept of co-production has been around for decades, but has in recent years experienced a revival. Research in a variety of disciplines has paid increasing attention to the role of citizens and the third sector in the provision of public services. The growth of interest in co-production during the past ten years provides important insights into, and at the same time poses important challenges for, public management.

More recently, exploring co-production has become increasingly topical for a broad range of academics with a focus on, and practitioners working with, public services and management. In 2006, the Public Management Review published a special issue on co-production called Co-Production. The Third Sector and the Delivery of Public Services, vol. 8(4). It was later reprinted by Routledge and made available in paperback (Brandsen and Pestoff 2006, 2009). Substantial work has also been undertaken by Alford in Australia and other Anglo-Saxon countries (2002, 2009). In 2012, a collection of papers was published in New Public Governance, the Third Sector and Co-Production, a volume published by Routledge (Pestoff et al. 2012). This renewed academic interest in the issue follows up and builds upon the work by early scholars like Parks and the Ostrom (1973, 1975). The importance of the topic of co-production is illustrated by the 2009 Nobel Prize in Economics awarded to the late Elinor Ostrom for her work on the analysis of common (public) goods and the role of users and their associations in producing such goods (Ostrom 2009).

The concept of co-production is at the crossroads between several academic disciplines, which makes it an increasingly targeted object of study by many scholars. Following previous work in this field, particularly in the tradition of Ostrom (1999), co-production can be defined as ‘…the mix of activities that both public service agents and citizens contribute to the provision of public services. The former are involved as professionals, or ‘regular producers’, while ‘citizen production’ is based on voluntary efforts by individuals and groups to enhance the quality and/or quantity of the services they use’ (Parks et al. 1981, 1999). In this widely accepted definition, the relevance of the concept of co-production becomes clear and perhaps more importantly several research challenges emerge, which may be addressed by scholars from different background.

Sociologists may be attracted to the voluntary element that is inherent to co-production. People deliberately choose to contribute time and effort in the production of services that were previously the responsibility of professional actors and organizations alone. What is different as opposed to classical volunteering is that co-production takes place within a context of professionalized service delivery and that it concerns services the people involved themselves use, i.e. not or not solely for the benefit of others. The IKEA model is an often-used parallel from the business world.

This bridges the gap between voluntary sector research and public management research. This has already been decreasing in recent work in the latter tradition, which acknowledges the increasingly fragmented and uncertain nature of contemporary public management, referred to as the New Public Governance paradigm (Osborne 2010). ‘It posits both a plural state where multiple interdependent actors contribute to the delivery of public services and a pluralist state, where multiple processes inform the public policy making system’ (Osborne 2006, p. 384). As such, the emphasis co-production puts on citizen involvement in the production of public services, makes it an element of this emerging paradigm: services are no longer simply delivered by professional and managerial staff in public agencies, but they are co-produced by users and communities. Traditional conceptions of service planning and management are, therefore, outdated and need to be revised to account for co-production as an integrating mechanism and an incentive for resource mobilization (Bovaird 2007).

In this article, we will summarize the state of the art in scientific research into co-production. The research of the past 10 years has advanced the debate considerably. There is now more evidence for the dynamics, benefits and drawbacks that had previously been defined only theoretically. Furthermore, increasingly, the research is comparative in nature, which we regard as an improvement in the quality of the work. But, there is still much room for improvement. Particularly, conceptual confusion remains a problem and the methodological diversity is still limited. Much of the research remains primarily descriptive. Also, we still lack a comprehensive theoretical and systematic empirically orientated understanding of what happens when citizens and/or the third sector are drawn into public service provision, and of the various aspects of co-production. In the remainder of this article, we will thus provide a systematic overview of recent research as a state of the art (‘knowledge’) and discuss the challenges for future co-production research (‘knowledge gaps’).

Taking a look at the recent research on co-production, we could systematize the results along three broad lines:

  1. 1.

    Under which circumstances do people co-produce: what are their motives and what ‘kind of’ people co-produce?

  2. 2.

    How does co-production work: what are the key variables that make it effective?

  3. 3.

    Does co-production actually lead to better service delivery?

The Concept of Co-production

Before we start with our overview, firstly, we have chosen here a note on the meaning of co-production. There is a lot of heterogeneity in the terms that are used in the literature under the broad umbrella of co-production. Government and its agencies that co-operate with citizens and private organisations in defining policies (the policy-making stages of the policy making) has been referred to in the literature as co-construction, co-policy planning or co-prioritization. Essentially, these concepts refer to non-governmental actors being involved in making policy decisions (e.g. via referenda, via budgeting, via hearings or via organizational input in drafting legislation like NPO’s advocacy or even representation in parliamentary hearings or committees) or in evaluating policy. This is what Brandsen and Pestoff have called ‘co-governance’ (2006). Government that co-operates with private actors in service delivery is ‘co-management’, meaning that non-governmental actors have a say in the design of the service, or put time or other resources (e.g. money, skills, expertise) in the delivery of public services.

In this theme issue, we necessarily had to focus and therefore adopted the narrower, classical interpretation of co-production as the involvement of individual citizens and groups in public service delivery. For work that covers a broader range of interpretations, we refer to our recent edited volume (2012).

Why Co-production? The Question of Motives

As noted before, co-production is different from classical volunteering in that it concerns services the volunteers use themselves. It is commonly assumed, especially by economists, that people co-produce because they have a material interest in doing so. This line of reasoning, inspired by public choice theory, claims that people are benefit maximizers who will only co-produce when benefits outweigh costs. These benefits can be very different (Alford 2009) and relate mainly to the so-called extrinsic rewards: people co-produce in return for a material, extrinsic reward, that compensates for the time and effort spent whilst co-producing. This reward may be monetary (e.g. a voucher in return for community service) or non-monetary (safer neighbourhood in return for being a member of the neighbourhood watch). This self-interest motivation has its limitations, however, for explaining why there is co-production. Even economists from public choice schools acknowledge that people can be driven by other values too, like altruism or sociality (Alford 2009). There are plenty of examples in everyday life that prove the existence of such human behaviour: donations to charities, signing petitions, volunteering, etc.

According to Alford (2009), different motives beyond self-interest exist for co-production. He notes that ‘…eliciting co-production is a matter of heightening the value that clients receive from the services by making more explicit their non-material aspects through intrinsic rewards, solidarity incentives or normative appeal’ (Alford in Pestoff 2012). Intrinsic rewards can be powerful motivators since people are not only solely motivated by self-interest but also by social values. This includes the enjoyment associated with interacting with other people, gaining their approval or avoiding their disapproval. Normative purposes are also important for motivating co-production, including values like participation, influence and democracy. Thus, there are three types of motivation—intrinsic, social and normative—in addition to material rewards that can elicit co-production. In order to prompt clients to co-produce, an organization must offer them something of material, social or normative value (ibid.).

In addition to people’s individual motivation to co-produce, other circumstances may hinder or facilitate co-production. According to Pestoff (2012) and Alford (2009), the ease of involvement and the willingness of individuals to participate in the co-production of services are important, re-inforcing factors for why people decide to co-produce. The question people ask themselves is how easy it is to get involved and why they should in the first place. The ease of citizens becoming involved will depend on several things, like the distance to the service provider, the information available to citizens about the service and its provision, etc. They are related to the time and effort required for citizens to become involved and might therefore be seen as the transaction costs of participation. If and when opportunities exist for motivated citizens to participate actively in the co-production of a service, lowering the transaction costs will make it easier for them to do so. However, the greater the effort required for citizens to become involved the less likely they will do so.

As Pestoff (2012, p. 24) states: ‘Citizens’ motivation to become involved as a co-producer will, in turn, depend on the importance or salience of the service provided. Is it a very important service for them, their family, loved-ones, a relative, a friend, or not? This will reflect how the service affects them, their life and life chances. Does it make a direct impact on their life and/or life chances or does it only have an indirect effect? If and when a person feels that a service is very important for them and/or their loved-ones or vital to their life chances, they will be more highly motivated to get involved in the co-production of services. It is, therefore, necessary to make a distinction between enduring and non-enduring services. Many social services belong to the former category, and therefore have an immediate impact on the life, life chances and quality of life of the persons and/or families receiving them. The importance and impact of such services guarantees high client interest in the development of such services, especially in service quality. Enduring social services include childcare or preschool services, basic and higher education, elder care, handicap care and housing as well as preventive and long-term health care. Users of such services are locked into them for a longer period of time and can therefore not normally rely on exit to provide them with influence or redress. The transaction costs of exit are often prohibitive (Pestoff 1998) so voice, rather than exit, provides clients with influence and redress’.

Pestoff continues (2012, p. 25) ‘that it is important, however, to understand that citizen involvement is more than just a question of facilitating greater citizen participation or developing techniques to motivate them. It is a combination of the ease of involvement and individual motivation. In other words, citizens are not like a ‘jack-in-the-box’, just waiting for someone to push a lever that will immediately release their energies and result in their engagement in social service co-production. They need to be motivated to do so, but the greater the effort required of them to overcome hurdles to participation, the greater their motivation must be. However, the greater the effort required of them to become involved in co-producing a public service, the more the service provided by a public agency must be both relevant and salient for them personally. Thus, less ease of involvement may thwart greater citizen participation, even in highly salient services, thereby limiting their participation to ad hoc, spontaneous and individual involvement’.

What Makes Co-production Effective? Questions of Organisation

A second broad theme in the current research on co-production deals with the conditions under which co-production takes place. The classical starting point for discussing the conditions under which effective co-production can take place is offered by the design principles for successfully managing common-pool resources (Ostrom 1990). These principles can be applied to discussions on effective co-production as co-production is about the involvement of users in the delivery of goods that are at the disposal of a group of entitled users (like safety in the neighbourhood, education in school, nurturing in pre-school child care, etc.). The main principles identified by Ostrom are

  • Defining the boundaries of the resource itself (neighbourhood in which safety is to be preserved, school in which education is provided) and defining the group of users.

  • Adapting the rules concerning the use and provisions to the local circumstances (who is entitled to what? Who contributes what?).

  • Letting co-producers be involved in the decision making, directly or via participation.

  • External authorities should restrain from being too involved with the right of communities to organize themselves.

  • Developing a (social) infrastructure for resolving conflicts between the actors involved.

These principles give us some direction for the development of processes in co-production initiatives and have been confirmed by other studies, for example by Brandsen and Helderman in their study of German housing cooperatives (2012, also in this issue), where people claim their right as a community to organize their housing themselves and develop systems and processes to do this effectively.

Subsequent work on co-production has looked more specifically at the organisational conditions that make it successful. Roughly, this can be divided into two types of work. One concerns work processes within organisations and how they affect co-production, whatever the type of organisation. The second examines what type of organisation (for-profit, public, third sector) is most apt at creating effective conditions for co-production.

Intra-organisational Conditions for Effective Co-production

Alford (2009) has examined how management activities and organisational processes can be reoriented to make better use of client co-production. The key is making clients the object of attention in the organizational process. Thus, understanding the clients’ needs is an essential condition for co-production. The organization must also question what they require from the clients. Hence, we have a two-sided coin with essential questions about reciprocal needs and expectations between client and organization (regular producer). From the organizational perspective, it is important that clients fully understand the value that the organization is seeking, hence the organization must clarify the value it is seeking to achieve: ‘what are we trying to do here?’ This question is about outcome or effect, rather than input, process or output. Answering this question will enable the client to understand the mission and purpose of the organization with which the client can identify and feel ‘appealed’. Also, having a clear focus on the organizational values will enable to better define in which parts of the production process and for what purposes different sources of co-production can be beneficial. In other words, co-production strategies can only be developed effectively when the ultimate outcomes are defined (and understood) clearly. Alford (2009, p. 206) gives the example of programs for the unemployed: where the goal is to get people to a job as fast as possible, then the strategies will focus on ‘work first’. If the goal is to get people to a stable and sustainable job, then the strategy will probably focus on ‘education first’. The kind of co-production then will also have to vary with the different strategies chosen. This brings up a second requirement from the perspective of the regular producer: analyzing the production process that leads to that particular value or outcome. The analysis of this public production process involves some necessary steps: drawing a chain of causality (identifying the factors that are likely to cause the outcome to be achieved), identifying key points in the chain (like bottlenecks, or issues affecting costs) and the people associated with these points, determining how to influence these people and developing a strategy to integrate the choices of co-producers and their methods.

The other side of the coin in managing effective co-production is to meet the clients’ needs. This requires, in the first place, knowing the clients’ needs and motivations to co-produce. Secondly, especially in a public sector context, managers of an organization always need to find a balance between adding value for service users or clients and adding value for the community or the general citizenry. Sometimes these values may conflict. Alford (2009, p. 210) gives the example of co-production in prisons: motivating inmates to co-produce may be an effective way to manage the prison and an innovative way to treat inmates, but may also come into conflict with how the public opinion thinks inmates should be treated. This balancing requires that organizations look for trade offs between public and private value so that both clients and the citizenry gain from the organization’s service offerings.

Finally, other important issues to take into account when managing for effective co-production include developing the right organizational structures that facilitate co-production. According to Jaworski and Kohli (1993, as cited in Alford 2009) such structures should be characterized by low centralization and high connectedness: e.g. managing relations with clients, autonomy in decision making for lower organizational levels to make judgments on the spot and coordinating answers to clients’ problems. Besides structure, organizational culture must also be carefully considered so that client focus and client engagement becomes institutionalized in the organisation, as a part of its culture.

This also touches upon the relationship between professional staff in the regular producer organization and the ‘non-professional’ users/clients involved. Professionals may contend that their own education and experience are more important than user involvement. If not considered seriously, these concerns may result in lower quality of services instead of equal or better quality service delivery, as Vamstad shows (2012; also in this issue). Related to this, there seems to be a constraint on the amount of substitution that can be undertaken between citizens and regular producers (Brudney and England 1983). Often, citizens lack training and skills to perform some tasks, and substituting paid personnel with volunteers means that some of the costs would be transferred to the co-producers themselves.

Porter (2012) points at some other important variables or conditions that should be fulfilled to make co-production effective: both regular producers and co-producers may need some skills in addition to their motivation to co-produce in the first place. In education, for example, students and families with better co-productive skills often receive a disproportionate share of education services. Thus, especially in a service delivery that relies on co-production to be effective (like education), the skills of co-producers and their families are essential. Often, these skills are associated with education levels and income levels of co-producers, like parents of pupils (Rosentraub and Sharp 1981). Similarly, regular producers (like teachers) also need some skills to make co-production effective. Secondly, both regular producers and co-producers must be willing to invest their time and money. In services that rely on large co-production by consumers, like education, the investments made by the co-producers are often not taken into account when financing and budgeting the service or when calculating costs. Still, investments by co-producers are often considerable in terms of time, money and often intangible efforts like support, encouragement, etc. Acknowledging these investments is essential in designing systems of effective co-production. Thirdly, (actors in) the networks in which regular producers and co-producers are embedded need to be supportive for co-production. Many services, or systems of service delivery, are embedded in networks of actors that constitute the so-called task environment. Again, in education, the teacher–pupil relationship is central, but surrounding this relationship are parents, peers, the community, etc. All these actors in the task environment also give input in the service delivery system. Porter (2012) paints a picture of a polycentric network structure between these actors with many nodes that are adjacent to the core process of co-production (in the case of education: the student/teacher nexus). The inputs that these nodes deliver to the core process may be supportive (e.g. parents stimulating pupils to do homework, local government issuing regulation that is supportive for establishing neighbourhood watch) or inhibitive (parents not being a good role model or police departments interfering too much with the practice of the neighbourhood watch).

Another variable that is under researched in relation to co-production is the new technologies. According to Meijer (2012, also in this issue), technology matters in an instrumental and institutional sense, especially in the information age. Firstly, from an instrumental point of view, technology is able to facilitate new practices of co-production: costs of large scale and dispersed action can be lowered and new media can make co-production more ‘social’ and ‘playful’. But secondly, from an institutional point of view, he claims that the added value of technology depends on the policy domain in case, on the institutional situation, and on the existence of citizen communities willing and able to link co-production with technology. For example, initiatives like TimeBanks (Cahn and Gray 2012) may largely benefit from new technology, as is the case with participatory budgeting: ICT supports co-production in the sense that co-production is made ‘easier’ for the potential co-producer, and in the sense that the coordination cost may be lower (co-production easier to organize). As Meijer shows, also in more complex cases of co-production, like public safety, ICT may enhance co-production. He gives the example of CitizenNet implemented in Dutch police departments: citizens give their address to the police and the police can contact these people whenever they need information about something that happened in the neighbourhood of the citizen (a thief running away in the area, a stolen car, somebody who is missing, etc.).

Other variables for effective co-production that have been reported in recent literature include integrative structures and building relational capital between the stakeholders (Brown et al. 2012), a high degree of organizational flexibility in participating organizations and a sense of shared responsibility for the provision of a new service (Schlappa 2012). Such factors are well known to enhance the effectiveness of networks of public service delivery, as documented in a large literature (Castells 1996; Agranoff 2007): trust, reciprocity and shared values and missions between the actors in the network.

Types of Organisations in Relation to Co-production

A second debate concerns the type of organisation that is most effective in achieving co-production. This discussion has particularly focused on the distinctions between third sector, public and for-profit organisations. Comparative work of this nature remains relatively rare, but there is some, which tends to emphasize the third sector’s role.

One study that looked into this is the TSFEPS Project that examined the relation between parent participation and the provision and governance in childcare in eight European countries (Pestoff 2006, 2008). From a comparative perspective, this study found different levels of parent participation in different countries and in different forms of provision: public, private for profit and private non-profit. It found that in non-profit service delivery, individual participation was best facilitated; an observation that was confirmed by Vamstad (2007) in his study of Swedish childcare. More than in for profit and public childcare initiatives, parents in non-profit childcare were able to co-produce and to assist in managing and governing the child care facility. This observation may say something about how professional organizations (regular producers) relate to service users (the assumed co-producers). The evidence suggests that in non-profit professional organizations, the conditions for engaging users are better, at least in the case of childcare. In the case of parent cooperatives in child care (where parents/users establish and organize the service themselves), this observation should not surprise, as co-production is endemic in cooperatives. The finding that co-production is most developed in parent cooperatives is even strengthened by the observation that users of public and for-profit child care in Sweden express a clear desire for more user influence, compared to users of cooperative child care. This shows that the desire for influence by users is best satisfied in non-profit (cooperative) child care.

Valuable as these studies are, the evidence is still limited. Also, several questions must be raised over the nature of the comparison. To begin with, not all third sector providers are equally able or willing to facilitate greater client participation in the provision of public financed services. They may more easily be able to embrace co-production than either public sector services or for-profit providers of the same kind, but whether they do or not depends in part on their own governance structures and the degree of democracy found in their internal decision making. Also, the outcome of any comparison between for-profit, third sector and public providers is likely to depend on the public administration regime in a sector and country. In Sweden, for example, most preschool services are provided in a traditional top-down public administrative fashion by the municipalities, whilst private for-profit preschool services seem inspired by ideas of greater consumer choice. By contrast, parent cooperatives promote both choice and voice and even require membership participation or co-production. This specific context may explain their growth in recent decades. However, it will take research in different public administration regimes to confirm this definitively.

What are the Effects of Co-production?

Another import research stream focuses on the effects of co-production on the (perceived) quality of services delivered, on the democratic level of service delivery and on the level of accountability towards many stakeholders when public services are co-produced. This is the least developed part of the research on co-production and one that definitely requires more attention in the future.

Efficiency and Quality of Service Delivery

Already in the 1980s, Warren et al. (1982, as cited in Pestoff 2006) claimed that co-production can lead to cost reductions, higher service quality and expanded opportunities for citizens to participate. Co-production then becomes an important means to enhance the quality and quantity of public services. Co-production also affects the nature of public service delivery. On the positive side, co-production can contribute to greater satisfaction of users to services. As argued by Pestoff (2012, also in this issue) and Calabro (2012), public service delivery in close cooperation with citizens and the third sector has the promise of delivering better public services in the eyes of a key stakeholder of public services, i.e. the citizens.

Although often stated, there is now actually some evidence that co-production may lead to better quality service delivery. Greater involvement of users in service delivery can lead to higher levels of satisfaction due to greater ‘moral ownership’ and tailoring of services to personal needs. Broadly speaking, it allows some of the benefits of market consumerism into the realm of public services. Vamstad (2012, also in this issue) found that co-operative child care in Sweden offers ‘better quality’ compared to professionalized municipal child care in the eyes of both clients and staff. According to Vamstad, co-production seems to be a promising method to reach the goals for public management that is not coming from the market-oriented literature originally intended for the private sector. Obviously, there needs to be much more work of this kind to draw any wide-ranging conclusions.

Democracy and Accountability

Cahn and Gray (2012) have noted that citizen co-production in the advancement of public goods and services has a rich history in the United States. The concept embraces a wide range of volunteering, but it also can lay claim to distinctive progeny stemming from the civil rights movement and the Johnson Administration’s War on Poverty. The statutory mandate ‘maximum feasible participation’ in the Economic Opportunity Act of 1964 sought to enfranchise the poor with both a voice and a role in the implementation of the programs initiated as a part of that effort. This idea continues today with the concept of Time Banking. Cahn and Gray provide numerous examples of co-production generated by Time Banking, which essentially developed its own version of co-production as a catalytic vehicle that takes ‘maximum feasible participation’ as a source of citizen empowerment. In a study by Calabro (2012), co-production is proposed as an alternative to privatization of public services. By developing new ways for citizens to participate in the production and provision of public services, a new relationship between citizen and state can be developed that is based on trust, greater ethical standards and accountability. His claim is based on the observation that that many acts of privatization of the 1980s and 1990s in Italy and Norway have only resulted in ‘partial’ privatization of public services formerly delivered by the state, but without introducing real market competition for these services. These partial privatization processes can be detrimental for the accountability of the service providers towards the citizens/clients. Therefore, co-production is proposed as an alternative.

However, there are, of course, also drawbacks observed in the literature. To begin with, co-production can strengthen insider/outsider dynamics, when this type of service provision is only accessible to specific social groups, either because these groups actively guard their own borders or because there are institutional mechanisms that discourage certain groups from engaging. This is shown by Brandsen and Helderman (2012, also in this issue): there is a gap between the rhetoric of housing cooperatives (on paper, anyone can join at a relatively low cost) and the reality (cooperatives are in practice rather closed systems). Another related issue—‘who is involved in co-production’—is still suffering from a lack of systematic evidence: we do not know much about the social background of those involved in co-production and the cases are indeed conflicting. A second and related drawback is that co-production appears to raise equity issues (Rosentraub and Sharp in Porter 2012): wealthier, better educated and non-minority citizens may be more willing and able to engage in co-production activities. If this is true, a large group of citizens is excluded from co-production and that is all the more problematic knowing that these groups often need the services produced the most. A third potential drawback of co-production is that the issue of ‘accountability’ may be problematic: who can the users hold accountable when the users themselves are part of the production process?

Shortcomings in the Current Research

Having provided an overview of the state of the art in the research on co-production about some key questions that are of interest, let us identify the main weaknesses of current research in co-production. To begin with, there are conceptual issues that need to be explored further: what are we talking about when discussing co-production, are we talking the same language? Moreover, there are methodological issues: what research strategies do we need and how do we measure what we want to measure?

Conceptual Issues

We showed throughout the article that co-production is a rather heterogenous umbrella concept that runs the risk of being used for many different purposes: is it about policy-making or policy implementing, is it about individuals or organisations that co-produce? As Pestoff (2012) shows, there are other complexities as well. For example, who is exactly the co-producer: the direct beneficiary himself or his relative? Or, who are the co-producing parties, both on the side of the regular producer (professional staff) and on the side of the co-producers (citizens, clients, TSO’s). Also, there is a certain difference between more ‘mundane’ and more ‘enduring’ co-production although borders are not always easy to draw. Examples of ‘mundane’ co-production are for example filing personal tax-returns or filling in postal codes on letters (Alford 2009). In the strict sense, this is co-production because the contribution of people (although a very small one) assures timely and effective delivery of letters and tax administration. Other types of co-production focus on more enduring services, and it requires greater effort over a longer period of time: child care, a neighbourhood watch and parents assisting at homework (education) are examples. This conceptual complexity urges researchers for carefully defining what they are researching and for carefully explaining how their research objects can be positioned in the broader literature on co-production. Conceptual rigour is important if we want to embark on research that is comparative and also for methodological clarity (what do we measure, and how?)

Methodological Issues

Most of the research on co-production is case study based. Only a few references are made to data collected in a large-N setting through surveys and so forth. Vamstad’s article in this issue is one exception. This, of course, limits the scope of the findings and should be compensated for in future research. If we could measure co-production in a large N setting, we could test hypotheses and theorize the determinants and effects of co-production, for example, trying to measure and compare the level and intensity of co-production in different settings. Furthermore, and partly as a result of case study designs, most authors use qualitative data. Perhaps this is not surprising, given the ongoing conceptual debates on what ‘co-production’ actually means. Conceptual clarity is a first prerequisite to be able to operationalize the concept of ‘co-production’ for reasons of quantitative measurement. Some authors have tried to quantify their data. Calabro (2012), for example, has measured the number of public service providers that apply codes of conduct in public service provision and how issues like corruption and interest-conflicts are avoided through the use of such codes and rules. This is, however, a measurement of a possible effect of co-production (ethical behaviour and accountability whilst delivering public services), and not a measurement of (the presence of) the ‘co-production’ itself.

On the whole, the research in the field has moved on. Whereas previous work was dominated by single case studies, there is now much more comparative work. Yet, it is also evident that the nature of the comparisons is so different that it limits the cumulative effect of the research. Therefore, future research efforts should bring together different scholars from different countries to design comparative research: the same research questions, the same concepts, operationalized and measured in the same way in different settings (countries and/or policy fields and/or services).

Directions for Future Research

How to move on? We already noted some issues in the co-production research that warrant further attention. However, for systematic advancement, it will be necessary to link the study of co-production more explicitly to general theories widely accepted in the social sciences. It is clear that the most parsimonious theorizing on co-production is still economic in nature and relates to the division between public, private, common pool and club goods, and to the ‘three failures’ theorizing closely linked to this classification of ‘goods’. Therefore, it may be interesting to explore links to other types of theories taking the theorizing on co-production out of the dominant economic models and bringing in competing theoretical frameworks that will enrich our understanding of the phenomenon.

To conclude this article, we will make some suggestions on the basis of the three issues discussed in this article (why, how and effects?), taking the overview of theories of government-non-profit relations that was published by Smith and Grönbjerg (2006) as our point of departure. They distinguish three broad frameworks for conceptualizing the relationship between government and non-profit sectors: (1) a demand/supply model; (2) a civil society/social movement model; and (3) a neo-institutional model. In each of the three models, further variations can be distinguished. Most of these theories have something to say about at least one of our three broad questions: Why co-production? How co-production? Effects of co-production?

Demand/Supply Theory

Demand/supply theories are mainly economic in nature and highlight citizen demand and resulting public service supply structures. For example, the market niche model mainly teach us something about when and why co-production is a valid alternative for other types of service delivery (e.g. private market or government). It relies on theories of market failure and government failure to explain why markets or governments are not the best options to deliver the service: as a result, the non-profit sector or even individual citizens may be in a better position to (co-)produce the service in case. Sources of failure that regular producers may be confronted with include, e.g., lack of trust with markets (Hansmann 1980), free riders, the governmental focus on the median voter (Weisbrod 1977), limits to the political system and political cost of failed initiatives (Douglas 1987). Moreover, non-profits or organized citizens in some cases can meet special niche demands because they have access to specific resources or supply structures (James 1987): voluntary contributions, moral entrepreneurs not motivated by personal rewards and the presence of a kind of common good that reflect the value of a group of individuals (community of like-minded individuals, Smith and Lipsky 1993).

The transaction model essentially explains co-production as an agreement between partners that, by cooperation, compensate for each other’s failures (Salamon 1987). There is a clear and direct incentive with all parties involved to co-produce: co-producers are able to deliver services that government wants or needs, but cannot deliver itself due to government failure. On the other hand, the government is able to compensate for the voluntary failure of the co-producer via the deliverance of resources that compensate for the lack of non-profit or co-producer’s resource, like public finance, and for the lack of professionalization on the side of the co-producer. The transaction model also questions the nature of the process of cooperation or co-production. For example, what is the nature of the transaction between the partners in co-production? Salamon (2002) and Smith (2002) argue that the transaction-relationship between government and non-profits is very diverse, and goes beyond exchanging financial resources. Moreover, there are costs and benefits in the transaction: the financial relationship with government may provide the co-producer monetary resources, but in the same vein this may threaten the co-producer’s independence and legitimacy. The relationship may provide the co-producer with opportunities to learn and to professionalize, but equally this requires effort and time on the side of the co-producer.

Civil Society Model

The civil society model is fundamentally based on the early writings of De Toqueville, who argued that voluntary associations are vital for a healthy society: protecting interests of free citizens against the temptation of government to restrict this freedom, and providing a vehicle to influence government policies. The prevalence of non-profit (or organized citizens) initiatives in public service delivery is a result of the interplay between demand, social capital and cooperative social networks. The image of the government—nonprofit relationship is one of tension, because it assumes that communities and associations should address social problems, rather than government: the so-called self-service society (Glazer 1989). In the civil society models, the idea of individual responsibility and obligation to the community is very present (e.g. Etzioni 1993), with individuals and non-profits that take up responsibility also playing a role in mobilizing demand for policy changes. Another idea in civil society models centers on social capital (Putnam 1993): voluntary associations are crucial in the building of that capital that is needed for making healthy democracies.

The importance of these models for the study of co-production is twofold: firstly, it can tell us something about the motives of coproduction. People or organizations engage in co-production because they value the opportunity for influencing government and its policies via co-production. The difference with classical Tocquevillian analysis is that it focuses less on advocacy through representative democracy than on the design of public services in accordance with the values of the concerned citizens (e.g. parents organizing pre-school child care because they are not satisfied with the services provided by regular child care). Secondly, it can tell us something about the effects of co-production, and more specifically the value of co-production for building social capital by making people responsible and by bringing people together.

Neo-institutional Theory

Neo-institutional theory explains the features and nature of public service delivery and the role of non-profits and organized citizens in this as a result of the institutional environment.

Looking at macro-institutional models (e.g. Salamon and Anheier’s social origins theory 1998), the question is whether co-production can be explained as the result of historical developments by which political institutions are shaped by social class in a given country. The interaction between social classes, and their relative power, as historically developed, determines the size, roles and shapes of the non-profit sector. Such theories are far wide of mainstream work on co-production, but they could offer a starting point for explaining variations between levels of co-production within countries. They point to contingencies (such as power distribution between actors, history of public service delivery institutions, allocation of state resources between institutions, etc.) that determine the way in which services are delivered. For instance, a hypothesis could be that in an environment with a strong non-profit sector, and with a tradition of co-production, the presence of co-produced public service delivery will be (historically) strong. It also raises awareness of taking into account contingencies that are specific for a country, or policy sector, when doing comparative research on co-production. We can think of the legal, political and institutional environment in a country.

A second approach in neo-institutional theory stems from the ideas on isomorphism (Powell and DiMaggio 1983). According to this line of reasoning, the reasons why people or organizations co-produce, or the processes along which co-production unfolds, will be determined by processes of coercion (co-production because contributions in public services are made compulsory), imitation (best practices are copied and translated to other settings) or socialization (co-production is the result of values that stress altruism, volunteering and solidarity). This too is a promising line of enquiry that could be further explored.

In sum, we can conclude that co-production research may be informed by a large diversity of theories. Economic theory (demand/supply) may help us in further explain why people co-produce and under which circumstances this may be an effective and efficient way to deliver public services. Moreover, transaction theories help us to focus on how the relation between partners in the co-production relation can be understood. Social capital theories can be helpful to tell us something why people co-produce (discontent with the more traditional ways public services are delivered) and can strengthen the discussions about the value of co-production in terms of social and societal added value (democracy, volunteering and creating social bonds). Finally, the strength of institutional theory is in the emphasis it puts on the interaction between service delivery arrangements and the environment. As such, it helps us in understanding how environmental features explain the motives for, and processes of, co-production.

Conclusion

In this article, we have outlined the main paths along which co-production research has proceeded: what are the motives, what are the processes and what are the effects of co-production? We have shown, by discussing some key-publications, that recently a lot of renewed academic interest in the subject has emerged. Notwithstanding the criticism that can and should be leveled against current research, significant steps forward have been made over the past few years. However, there are important challenges left ahead. The growing interest in co-production, the number of services that refer to co-production as the way to public services of higher quality and the assumed advantages of citizen involvement, force us to continue to take a closer and critical look at this phenomenon.

The challenges in doing so are twofold. Firstly, the field is in need of empirical and comparative research that tests the assumptions that are attributed to co-production, like the generation of social capital, new relations and cooperative behaviour, greater quality of public services and reduced costs and greater benefits for those involved. The research on co-production would also benefit from greater methodological diversity (specifically more quantitative comparative work) more diverse use of general theories and further conceptual clarification. Looking at the work that is currently going on, the future looks bright.