‘I do not want to be one of her favourites’. Emotional display and the co-production of frontline care services

ABSTRACT

Co-production has been introduced into many public services, reshaping traditional roles of frontline staff and service users. This study investigate how co-production processes interact with norms of emotional display set by staff. A continuum of staff-user relationships is developed, highlighting variations between display rules which set norms of emotional closeness versus distance. This continuum is used empirically to explore relationships as frontline care services introduce co-production initiatives. Data from a Danish case illustrate that emotional closeness/distance can influence the effectiveness and sustainability of co-production. This finding contributes to understanding of the factors which underpin enduring co-production in public services.

KEYWORDS:

• Co-production
• social care
• emotional labour
• professionalism
• boundaries

Introduction – rethinking the staff-user relationship

Although the death of New Public Management (NPM) is much disputed, there has been a shift in the aspirations of many European states to look to alternative models of public service. One influential model has been New Public Governance (NPG), an institutional logic which differs from NPM-type approaches in its conceptions of the relationship between state, market and civil society (Osborne 2010; Wiesel and Modell 2014). The concept of co-production is a key element of NPG and has gained interest worldwide amongst both academics and governments (Verschuere, Brandsen, and Pestoff 2012; Osborne, Radnor, and Strokosch 2016). Although co-production has many definitions, all share the key insight that it remakes the relationship between service users and public service providers (Pestoff 2019). Co-production is seen as having potential to improve the quality of services and deliver broader economic and social benefits (OECD 2011; Bovaird and Loeffler 2012; Jaspers and Steen 2019; Brix, Krogstrup, and Mortensen 2020). There is also a normative drive here: as governments increasingly focus on the wellbeing of their populations rather than narrower efficiency-oriented goals, co-production at the frontline is recognized as a key component of self-efficacy and wellbeing (Needham 2008; Loeffler and Bovaird 2018). However some aspects of the transformed professional practice and staff-user relationships necessitated by co-production are underdeveloped, both normatively and empirically (Park 2019). Bartels and Turnbull (2020) have recently affirmed the significance of relationality in public administration studies and have suggested that a network of interactions, interdependence and relationships act as a main unit of analysis in relational public administration. Our article contributes to the growing literature on relational approaches within administrative practices (Bartels and Turnbull 2020; Ropes and de Boer 2021) by focusing on the following research question: if co-production prescribes new relationships between frontline staff and service users, how does this interact with the norms of emotional display within public services?

In the first section of the article the concept of co-production is discussed to develop a theoretical description of the norms shaping the relationships between staff and service users. Second, literature on the display rules of emotional closeness/distance and emotional boundary work is reviewed. A continuum of emotional closeness is proposed, highlighting the potential for co-production to pull either towards the closeness or distance end of the continuum. Emotional closeness/distance is then considered in relation to a Danish example in which two disability services are implementing co-production. The area of disabilities was chosen for this study as relationships between staff and service users are essential for people with cognitive and learning disabilities’ life, independence and general wellbeing (Robinson et al. 2020). Despite this, the impact of co-production on the relationships between frontline staff and people with disabilities remains under-explored (Robinson et al. 2020). This study addresses this gap and explores how the emotional display rules set by staff may affect co-production processes and the desired goal of users’ independence.

Co-production and staff-user relationships

The concept of co-production was originally developed by Elinor Ostrom and her research group in the 1970s, although it took a long time to reach the public policy mainstream. Ostrom described co-production as the potential relationship that could exist between the ‘regular’ producer and the ‘clients’ who want to be transformed into safer, better educated or healthier persons (Ostrom 1996). Co-production’s emphasis on collaborative arrangements between staff and service users posited a strong challenge to the Weberian norms of bureaucracy which had been influential in postwar European and Anglo-Saxon public administration systems. Sometimes described as ‘Old Public Administration’ (OPA) to draw out the contrast with later approaches, this ideal type frames service users as passive clients of standardized public services (Wiesel and Modell 2014; Torfing, Sorensen, and Roiseland 2016). NPM, an alternative ideal type which gained traction in some countries from the late 1970s, reframed the government-citizen relationship on quasi-market lines in which service users became customers (Osborne 2010; Runya, Qigui, and Wei 2015; Krogstrup and Brix 2019). Like the early client model, there remained an emphasis on role differentiation and standardized interactions with service users (Needham 2007). Citizens had to ‘co-produce’ through, for example, taking medicines as prescribed, but we can see this as a minimal form of co-production based on compliance (Ewert and Evers 2012).

More recently, NPG-type approaches have presented a different norm of government-citizen relationships and a re-engagement with Ostrom’s work (Bovaird et al. 2021; Osborne, Radnor, and Strokosch 2016). NPG-type approaches draw their insights from social network theory and invoke the premise that multiple interdependent actors, including service users, contribute to public service delivery (Bianchi, Nasi, and Rivenbark 2021; Sørensen and Torfing 2021; Osborne 2010; Thomas 2013; McGann, Wells, and Blomkamp 2019). This implies a ‘radical reinterpretation’ of the user-staff relationship, which changes from a top-down, one-directional delivery model to a more collaborative relationship based on user empowerment (Radnor et al. 2013; Steen and Tuurnas 2018; Needham 2007). Evers and Ewert (2021, 136) have defined this form of co-production as a ‘new service and governance mode’, in which trust, relational capacities and relational contracts become a more explicit governance mechanism, at least in some services (Osborne 2006; Bartels and Turnbull 2020; Berg and Dahl 2020).In this article co-production is understood as active direct input by individual service users, which shapes the service that they personally receive (Brandsen and Honingh 2015). Following Osborne, Radnor, and Strokosch’s (2016, 640) definition of co-production, the involvement of public service users can be in the design, delivery and/or evaluation of public services. This also means that co-production is understood as a notion that ‘refers to exchange relationships that include several dimensions of interaction (e.g. dialogue, practical matters and cooperation) … ’ (Ewert and Evers 2012, 61). Thus, co-production enables the service user to shape the service by contributing knowledge, resources and ideas (Trischler and Scott 2016; Eriksson 2019; Osborne and Strokosch 2013; Sorrentino, Sicilia, and Howlett 2018). It increases ‘personal autonomy and service-user self-determination’ (Flemig and Osborne 2019, 1). Hence, frontline worker’s ability to create ‘meaningful power-sharing’, seeing the service users as individuals and foregrounding their strengths and resources, becomes essential to enable co-production processes (Tuurnas 2021; Needham and Carr 2009; Steen and Tuurnas 2018).

Despite co-production’s popularity, public service organizations have found it hard to implement and maintain co-production processes over time (Pestoff 2014; Bovaird 2012; Ness et al. 2014; Van Eijk 2017; Osborne, Nasi, and Powell 2021). Failed implementation is explained by various barriers such as practical/technical issues, lack of resources, and system-related and social-cultural aspects (McKenna 2021; Van Eijk and Gascó 2018). Studies have shown that changing the nature of staff-users relationships can be overwhelming and painful for the parties involved, causing, for example, professionals to cling to their old role perceptions (Torfing, Sorensen, and Roiseland 2016). These insights underline the need for further investigation into cultural and emotional barriers to co-production.

Discussions of barriers to co-production have also failed to explore whether similar or different dynamics are at work in the different phases of co-production: design, delivery and evaluation. Gheduzzi et al. (2021, 4), note that ‘Barriers to co-production are usually reported in broad terms without highlighting the process by which negative effects are achieved’. As a result it is not clear if the barriers causing failures in co-design (a dialogical phase) are the same as in co-delivery (a more outcome oriented phase). In this article, we study both co-design and co-delivery processes and investigate if different norms of emotional display impact on those processes. Below we set out two contrasting norms in relation to emotional display rules set by frontline staff in public services and then consider how these influence co-productive ways of working.

Emotional display rules and co-production

Emotion management is a key element of frontline work, ensuring that staff adhere to appropriate display rules in their own emotions and in managing the emotions of others (Ekman 1984; Morris and Feldman 1996; Hochschild 2012; Guy, Newman, and Mastracci 2014). If co-production is about enhancing trust and equalizing power relationships then it is likely to be influenced by as well as to influence existing display rules (Brotheridge and Grandey 2002; Ekman and Friesen 1971). A concept developed within the emotional labour literature, display rules are ‘the standards of behaviour that indicate not only which emotions are appropriate in a given situation but also how these emotions should be conveyed or publicly expressed’ (Morris and Feldman 1996, 988). As Humphrey notes, such rules ‘increase predictability and make service interactions go more smoothly; when display rules conform to normal social expectations, both parties (service agents and customers) understand what emotions should be displayed and how they should treat each other’ (Humphrey 2013, 80).

The ideal types of public management set out earlier invoke different types of display rules. Weberian bureaucracy valorizes depersonalized relationships as a way to overcome the limitations of clan and kinship types of support and to eliminate the personal, informal aspects of public organizations (Ouchi 1980; Leitner 2003; Bartels 2013, 470; Meier, Mastracci, and Wilson 2006). Arguably, the welfare state offered an institutional form of this model, tackling the inadequacies of family care and giving primacy to the judgment of the detached professional (Esping-Andersen 1999). There were efficiency arguments at work here, but also a commitment to ethical practice which eschewed nepotism and particularism (Abbott 2014). Gallos highlights the extent to which distance has been seen as a key competence of frontline professionals: ‘They need clear boundaries to sustain objectivity, protect themselves from the stress of the work, and nurture essential autonomy in others’ (2013, 47). The emphasis in this type of display rule is on maintaining boundaries, non-reciprocal support and staying in role.

NPM-type approaches steered public servants towards more market-customer relationships in some services (Needham 2007; Clarke et al. 2007). In this ‘customer is always right’ model, citizens were no longer required to be passive and grateful: citizens charters and minimum service standards indicated what was required and how to complain if that was not received. This more transactional version of public services moved away from the passive client model, but seemed to further intensify emotional disconnection between professionals and service users. Himmelweit writes of, ‘the emotional servicing of people who remain strangers’ in highly commodified forms of public service (1999, 35).

Of course, the notion of the detached, impartial professional, discharging the will of the state at the street-level, has long been challenged (Lipsky 1980; Maynard-Moody and Musheno 2003). As Gofen (2013) points out in a review of the deviance literature, street-level bureaucrats depart from formal rules in lots of different ways, including in developing more particularistic relationships with service users. The response of some of the literature is to identify mechanisms through which to bring bureaucrats back into line with norms of distance and formality (Gofen 2013). However in the literature on NPG there is increased awareness that relational capital is a key governance mechanism, rather than a dysfunction to be designed out (Osborne 2006; Bartels and Turnbull 2020). This links to a broader literature which locates ethical professional conduct in a more emotionally authentic presentation of the self (Bolton 2000; Baines 2011), Hochschild highlights the extent to which we prize authenticity: ‘as a culture, we have begun to place an unprecedented value on spontaneous ”natural’ feeling”'(Hochschild 2012, 190). This has found expression in the ‘bring yourself to work’ movement, and has been seen as one way of surfacing the hidden emotion work that is often done in organizations, particularly by women (Berg 2002; Portas 2018; Cottam 2018).

This more emotionally authentic set of display rules links to modes of public service in which caring about (as an affective disposition) sits alongside care as a type of service (Fine 2015). Whereas caring about is an inherently emotional state, the provision of care may or may not include an affective element. Writing about different models of care, Hochschild (1995) has drawn a distinction between ‘cold’ and ‘warm’ models. She sees institutional models of care as ‘cold’, emphasizing the control of the institution and the lack of caring about; care becomes ‘warm’ as the family becomes more involved. However later writers have identified scope for ‘warmth’ (or caring about) within institutional services. Tronto (2013) points out that family care isn’t always ‘warm’ and that there is scope to create ‘caring institutions’. In her study of care workers, Johnson (2015) found warmth of emotion towards care home residents. She cites one of the care workers she interviewed as saying: ‘if you don’t feel emotional you don’t care and then you’re not a carer’ (Johnson 2015, 122).

There are benefits to staff in these ‘warmer’ display rules through a fit with the caring mission of public service work and avoidance of dissonance and burnout (Guy, Newman, and Mastracci 2014; Breslin and Wood 2016; Farr-Wharton et al. 2021). The promotion of more emotionally authentic relationships is also seen as improving user outcomes – for example by making it easier to build trust when working with people who are suspicious of state intervention (Cottam 2018), or by helping to connect to people affected by grief (Bolton 2000). In examples of these more emotionally authentic display rules, frontline staff may hug citizens, socialize with them beyond work requirements; and use terms associated with love and family (Breslin and Wood 2016; Tanner, Ward, and Ray 2018). These forms of display can be tailored to particular individuals, rather than applying the same mode of interaction to all. It may accord with the display rules of friendship, in which staff describe users as friends, and take part in non-mandated activities with the users. Alternatively, more familial forms of display may be on show, associated with a high degree of support and closeness, e.g. physical touch, reciprocity of emotional support, discussion of love. For example, for people with disabilities who need residential care, England’s ‘Shared Lives’ initiative enables people with a disability to live with a family in a home setting, and has been described by its chief executive as ‘one of the few models that allows people to talk about love’ (Fox 2018, 22).

However, the ‘warmer’ forms of emotionally display, do run counter to elements of public service work. Emotional warmth can be in tension with the techno-rationality of caring as a professional competence or set of skills (Johnson 2015, 113; Meagher 2006). It can be hugely effortful for staff (Hochschild 2012), leaving them potentially emotionally vulnerable themselves (Berg 2002; Tanner 2020). It can be exploited by employers, presented ‘not only as natural but as naturally remunerated in the same moral currency of hugs and thank-yous’ to compensate for low pay (Johnson 2015, 117).

In Figure 1 we set out a continuum of emotional closeness, which distinguishes between on the one hand authentic/naturalistic relationships with informal and personalized display rules, and on the other hand distant/professional relationships with formal and depersonalized display rules. The informal/personalized end of the continuum is characterized by relaxed boundaries such that professional display rules adapt to accommodate norms closer to friend or family relationships. The other end of the continuum is characterized by greater emotional distance in which professionals maintain formal and depersonalized forms of display.

Figure 1. A continuum of emotional display rules.

We might expect that co-production aligns best with the more relational and emotionally authentic forms of display associated with the friend and family norms. It requires a weakening in the privileging of the professional perspective and a more reciprocal form of engagement in which power is shared and outcomes are agreed together (Kirkegaard and Andersen 2018; Bartels 2013). This may foster more authentic encounters, in which the conventional rules of professional display (restraint and distance) are removed. Display rules may be tailored to the person rather than applying to a whole group.

However, it is not necessarily the case that co-production best fits a context of informal and personalized display rules. Co-production might be undermined by personalized relationships with professionals if these have created an over-protective ‘gift’ model (Duffy 2010) and crowded out other forms of support. Co-production is not just about how services users engage with the state, but also how they develop user autonomy and positive risk taking (Bazemore and Erbe 2003; Jorm 2012). The Danish National Board of Social Services, for example, has launched a project aiming at improving disabled citizens’ quality of life by increasing and building peer relationships (The Danish National Board of Social Services 2014). In this account, co-production allows bureaucrats to take back distance as frontline staff roles are reframed as complementary and supportive to the empowered service users (Nederhand and Van Meerkerk 2017). Professionals thereby support service users in the development of more naturalistic relationships, ending the ‘crowding out’ by the state (Entwistle et al. 2007).

Below we use the empirical part of the article to investigate how norms of emotional display rules set by frontline staff in public services intersect with attempts to move towards more co-productive ways of working. In focusing on how the paid worker interacts with the person being supported, we recognize this as an over-simplifying binary. Within this relational space we might also find families, managers and volunteers. Here we focus on the dyad of frontline worker and service user as a way of highlighting some of the dilemmas and issues of co-productive relationships. We recognize, following Kirkegaard and Andersen (2018, 830), that it is through the frontline encounters that ‘categories such as professional [and] user’ are brought into being.

Methodology and setting

We draw on a Danish case study of services for people with learning disabilities. Care services, with their intimate connection to people’s daily lives and bodies, offer a well-developed case for consideration of the dynamic relationship between emotional norms and co-production (Robinson et al. 2020; Pestoff 2019). Care services have long been the site for debates about professionals’ emotional boundaries, given the intensity of support, the power imbalances between staff and users, and the potential for abuse (Mik-Meyer and Villadsen 2013; Kirkebæk 2008; Williams and Caley 2020). However, even in the care literature, there has been a lack of focus on how strong pressures towards co-production are shifting emotional display rules.

The Danish context was chosen as co-production is high on the strategic agenda of Danish municipalities and different national and local initiatives have been launched to support co-production processes (Krogstrup 2017; Tortzen 2018; Reiermann 2017). We, therefore, see this setting as more mature than other potential contexts in its embedding of co-production, offering insights into practices that may be more incipient elsewhere (Mortensen 2020). Aalborg Municipality (the third largest municipality of Denmark, with 213,589 inhabitants, located in the northern part of the country) is currently implementing an organizational change moving towards a higher degree of co-production (Mortensen 2020; Department of Care for the Elderly and Disabled Aalborg Municipality n.d.). The Department of Care for the Elderly and Disabled articulates this change as ‘leaving the care regime’ and argues, ‘If you are disabled, people around you often help and nurse you a lot. We have done that as professionals as well, but we have made a cultural turn [towards self-determination and independence]’ (Danish Design Award 2018, 1). The Department argues that this cultural turn requires supporting service users with a focus on their individual definition of quality of life, instead of support based on professional knowledge, intentions and assumptions (Danish Design Award 2018; Klinge et al. 2019).

A qualitative case study

The study draws on a case study with two supported housing services for people with learning disabilities in Aalborg, Denmark. Case 1 is a supported housing complex with 30 users. Case 2 is supported housing with 11 service users. In both cases, the service users rent their own two bedroom apartment and share a common area with a kitchen and living room. Both cases can be defined as long-term services with ongoing interaction between service users and frontline staff. This implies a high degree of interdependency within these services and that they have an immediate impact on the quality of life of the person receiving them (Pestoff 2019). The ‘opportunity space for co-production’ is a feature of the ‘mandatory conditions’ and ‘local conditions’ (Mortensen, Brix, and Krogstrup 2021) and in the following it will be used to explore similarities and differences between the cases. The cases have the same mandatory conditions, as they are within the same welfare sector and municipality, operate within the same legislation, and can be characterized as major, enduring services (Pestoff 2019). The workers in both cases are professionals with a Bachelor in Social Education. Local conditions cover, for example, the dominating institutional logic, operational priorities of resources, and cultural elements and social structures (Lipsky 2010; Berger and Luckmann 1991; Eriksson 2019). Again, the two cases have many similarities, as they provide similar services, are part of the same strategic direction, are given the same autonomy, and provide services to people with learning disabilities. There are some minor variations in the cases’ local conditions, which may affect the user-staff relationships. First, there is a variation in the duration of the relationship between staff and service users. Case 1 is a newly established unit, and most of the service users were settling in as the data collection processes started (in the summer of 2017), while the user-staff relationships in case 2 have existed for several years. Second, the cases also differ in the levels of reliance that service users have on staff. The service users in case 1 have a high degree of reliance, whereas in case 2 the service users are more self-supported and rely less on the staff for daily living activities. Due to the similarities of the cases, we expect similar rather than contrasting results (Yin 2014), although we do recognize that the local conditions influencing the interaction between staff may be relevant to emotional display, for example, emotional closeness may develop over time.

Ethical considerations

The first author was part of and partly funded by the Danish ‘Industrial PhD programme’ (Innovation Fund Denmark) and employed as an internal investigator at Aalborg Municipality and part of the secretariat at Aalborg Municipality’s Department of Care for the Elderly and Disabled. The author did not work in the services where the study was performed as she only undertook research tasks related to the PhD project in accordance with Innovation Fund Denmark guidelines. The central management of the Department of Care for the Elderly and Disabled at Aalborg Municipality gave ethical approval for the study. The individual case study design was developed in collaboration with the local managers, who were also responsible for informing staff, service users and guardians regarding the purpose and use of the research. The service users or their guardians gave their consent to participate in the project and were informed how to withdraw their data in compliance with the GDPR legislation and ethical practice.

Data collection and analysis

This study draws on qualitative data (observations and focus groups) collected by the first author in the two cases from May 2017 to February 2019 (See Table 1 for more details). Participant observation was used to gain insight into everyday practice and interactions between service users and frontline staff in the cases. Field notes were written down during the observation, and when possible the author recorded conversations and took pictures of different situations. All notes, recordings and pictures were combined and written up later the same day to produce extensive notes. The data from the observations were complemented by semi-structured focus groups with frontline staff and local managers. This was supplemented by more observations e.g. at the housing complex and at educational activities and seminars for the staff. The focus groups explored the nature of the relationship between frontline staff and service users, the everyday life at the housing complex, along with views and attitudes related to co-production and current service delivery.

Table 1. Summary of the fieldwork.

Case 1	Case 2
May-October 2017
About 30 hours of fieldwork at the supported housing 8,811 words of field notes and 25 pictures. 1 video clip and 2 audio recordings	About 30 hours of fieldwork at the supported housing 10,758 words of field notes and 41 pictures. 2 audio recordings
November 2017-September 2018
	14 hours of fieldwork at the supported housing and 11 hours of fieldwork taking part in educational activities and internal seminars for the frontline staff 3,089 words of field notes and 14 pictures. Detailed notes or audio-recorded
3 focus group interviews with the frontline agency (with 5–6 participants) (11 hours in total) (Audio recorded and documented through pictures)	3 focus group interviews with the frontline agency (with 6 participants) (9 hours in total) (Audio recorded and documented through pictures)
July 2018-February 2019
1 focus group with the frontline agency (with 4 participants)(1 hour) (audio-recorded)	1 focus group with the frontline agency (with 5 participants)(1.5 hour) (audio-recorded)
1-hour follow-up meeting with managers	4 hours of follow-up meetings with managers
3.5 hours of fieldwork at the supported housing

Initially, the first author made general notes and reviewed the empirical findings by listening to the audio files, reading through the transcriptions and observational data. These notes and empirical summaries were used to explore similarities and differences between the two cases. One of the differences was in what frontline staff considered appropriate user-staff relationships. It was here, in collaboration with the second author, that the relationship between co-production and emotional labour became a focus. The second author conducted the literature review on emotional labour in public services with a particular focus on varying degrees of emotional closeness and distance. Incorporating theory after data collection can be described as a transition from the inductive to the abductive, in which ‘data and existing theory are now considered in tandem’ (Gioia, Corley, and Hamilton 2013, 21). In collaboration, the authors developed the continuum of emotional display rules as a more general theoretical framing of the user-staff relationship, providing an anchor for the data analysis. Next, the data was coded based on two ‘parent nodes’ proceeding from the emotional labour literature. Each parent node houses different child nodes, which were generated continuously through the coding process. Table 2 summarizes the results of this coding process, along with examples from the data. Examples and quotes from the transcripts and observation used in the analysis have been anonymized with data labels, and translated from Danish by the first author.

Table 2. Child nodes and examples within the two parent nodes: Informal and personalized relationships and Formal and depersonalized relationships.

Informal and personalized relationships
Node name	Examples from observation and focus groups
Staff emphasize closeness and personalized relations to the service users as the foundation for their work	A. Staff emphasizing that there is a need for trust and reliance between staff and users B. Staff stating that it is essential for their work that the users feel secure and comfortable with them C. Staff highlighting confidentiality and familiarity as crucial elements of the staff-user relationship
Staff emphasize that they need to care and create safety for the users	A. Staff defining their job as guiding and caring for the service users B. Staff stating that it is their role to protect the service users from society and help them communicate with public actors and society in general C. Staff saying that the service users depend on them and need them to have a functional everyday life. Defining themselves as the ‘scaffolding’ in the service users’ lives and labelling themselves the as the users’ ‘lifeline’
Warm and friendly communication between service users and staff	A. Staff and service users teasing each other B. Staff and service users joking around C. Service users calling staff sweetie, ‘mom’, or other related terms
Physical contact between service users and staff	A. Service users hugging staff B. Service users asking staff for nuzzling
Staff comforting service users	A. Staff reassuring and calming service users
Cozy times between staff and service users	A. Drinking tea and coffee together B. Playing ball or board games together C. Watching TV together D. Watching videos on a service users phone E. Eating dinner together F. Going on walks and swimming together
Service users preferring the company of staff compared to other users	A. Service users choosing to be around the staff, rather than the other users
Structures and routines allowing staff and service users to have one-on-one time	A. Staff doing activities alone with service users (instead of in groups) e.g. going to the store, going on walks or swimming
Formal and depersonalized relationships
Node name	Examples from observation and focus groups
Staff emphasize boundaries and professional relation to service users	A. Staff did not want service users to differentiate between staff e.g. having a favourite staff member B. Staff seeing it as their task to teach service users that there is a difference between friends and them as staff – and that they are not their friends C. Staff believing it is inappropriate to have emotional involvement between service users and staff
Separation of cozy times (service users) and practical tasks (staff)	A. Staff doing practical tasks (e.g. cooking), while service users watch TV together, drinking tea or coffee etc. B. Staff choosing not to include or ask service users to help with practical tasks that they are doing C. Staff setting up and/or encouraging service users to do activities together or socialize without participating themselves
Formal and depersonalized relationships
Node name	Examples from observation and focus groups
Minimal interaction between staff and service users	A. Little or no communication between staff and service users during activities e.g. preparing food or lunch packet B. Staff not prioritizing spending time with service users e.g. choosing to do other tasks or talking to staff (rather than service users) about non-work-related topics C. Service users retreating from common areas and social situations (staff also might encourage them to do so) e.g. going to their apartment to watch TV, listen to music or eat dinner
Structures and routines fostering little one-on-one time between service user and staff	A. Little or no one-on-one time between staff, e.g. not driving in a car with a service user, taking the service users shopping in groups rather than individually B. Deliveries of pre-cooked meals for service users, which means that the service users and staff did not cook or eat dinner together C. New initiatives relying on volunteers (rather than staff) to do activities with service users e.g. go swimming or to concerts

Results: display rules within the empirical cases

Based on the coding and data analysis, it was found that examples of ‘Formal and depersonalized relationships’ and ‘Informal and personalised relationships’ were present in both cases. However, it was also concluded that each case overall was positioned towards one of the ends of the continuum. Case 1 had a majority of coding with the ‘Formal and depersonalized relationships’ node, while case 2 had most examples within the ‘Informal and personalised relationships’ node. A list of the nodes can be found in Table 2, along with examples.

Comparing emotional display in the two cases

In case 1, staff wanted to keep the relationship with users towards the ‘distance’ pole of the continuum. One of the staff members made the point that the staff should not ‘[…] be too emotionally involved, because it creates an inappropriate relation [to the service users]’ (Male employee in case 1, Focus group 2018). The argument was that the service users should have the same relationship with all of the staff and that nobody should be ‘too close’, because it could potentially complicate their work and the service users would be devastated if a staff member stopped working at the housing complex. The employee describes it this way:

I do not want to be one of her favourites [referring to a specific service user]. I want to be one of them who help her in her everyday life, but not a favourite. I want it to be good when I enter, but I also want it to be good when I leave again and another employee comes. Because that is how it is going to be their whole life … In the beginning she was very fond of me, but I withdrew on purpose, so I would not become her favorite. It is too hard for the service users. We have to work on this, so nobody becomes favorites (Male employee in case 1, Focus group 2018).

During the focus group the other staff agreed with this statement. Similarly, another worker also argued that it is their job to create awareness and help the service users to realize that there is a difference between their friends and the staff working at the housing complex or other actors helping them e.g. their drivers (Female employee in case 1, Focus group 2018). During the observations it was also noted that staff would withhold from socializing too much with the service users, for example by withdrawing from social activities such as eating dinner or watching TV. Also in case 1, structural routines discouraging one-on-one time between staff and service users had been adopted. This would, for example, guide frontline staff to take the service users shopping in groups rather than individually (which was not found in case 2). It can therefore be argued in case 1 that staff understanding of the appropriate display rules in their relationship with the service users fell within norms of distance and formality.

In contrast, the user-staff relationships in case 2 pulled towards the ‘closeness’ half of the continuum. The frontline staff in case 2 used words like trust and care to describe the relationship to the service users and argued that is part of their job to care for the service users. As one female employee put it: ‘It is difficult to be in our job if you are not a caring person. It is part of our job to care for other human beings. You are not really in the job, if you do not have care in you’ (Female employee in case 2, Focus group 2018). The frontline staff also felt that they were some of the service users’ network and ‘all they had’, as some of them had no regular contact with their family. The staff saw it as their responsibility to help the service users in many different parts of their lives, because they feared if they would not help them nobody else would. This led them to take on tasks such as going Christmas shopping, taking care of the service users’ pets, and decorating their apartments (Observation and focus groups in case 2, 2017 and 2018). This implies the frontline staff had a close bond and felt responsible for the service users beyond what is usually expected of staff in supported housing. In the observations the author also noted that the staff engaged in warm and friendly communication and would discuss personal issues with the service users e.g. the service users’ birth control options, relationship troubles with boyfriend/girlfriend etc. This was quite different to the formality observed in case 1.

Overall then the frontline staff in case 1 expressed a preference for an emotionally distanced, depersonalized relationship with the service users with an emphasis on boundaries, while in case 2 the staff spoke of a high degree of care and emotional closeness with service users, placing them, in the ‘closeness’ half of the continuum. We consider these to be ‘critical cases’ in terms of yielding contrasting insights into emotional display rules (De Vaus 2001; Eisenhardt 1989). In the section below we discuss how these different norms intersect with the requirement that the services become more co-productive.

Emotional display and co-production

At Aalborg Municipal Department of Care for the Elderly and Disabled, co-production and ‘collaboration with service users’ is a key commitment. There is a ‘letter of freedom’ which gives autonomy to the frontline staff to act upon the service users’ aspirations and their own initiative (Department of Care for the Elderly and Disabled Aalborg Municipality n.d.). Frontline staff are encouraged to co-design the services based on ‘the service users’ dreams and choice of lifestyle’ (Department of Care for the Elderly and Disabled Aalborg Municipality n.d.). When a service user is referred to/granted supported housing (or other public services) they get ‘provision goals’ and sub-goals, e.g. building social relations, becoming more independent in getting ready for work, personal hygiene etc. The service delivery needs to be provided within this framing, however in practice there will be a high degree of co-design and autonomy to decide how co-production is performed. As the Department puts it, ‘Self-determination and independence (or the feeling of this) are goals people reach, when we investigate their dreams in everyday life‘ (Danish Design Award 2018, 1, see also Klinge et al. 2019).

During 2018, part way through the fieldwork, both case sites began a co-production initiative, which progressed differently in the two cases. In case 1 the frontline staff wished to co-produce a series of ‘Friday afternoon get-togethers’ with the service users, where the users in collaboration with staff would design and be responsible for these events. At case 2 the initiative aimed at co-producing the daily dinner cooking in the shared kitchen at the housing complex, where the service users had a habit of taking a passive role. Both co-production initiatives were launched within the existing service delivery and affected the daily interaction between service users and staff. Below the implementation in the two settings is described, alongside a discussion of how it intersects with the different norms of emotional display found at the sites.

In case 1 (emotionally distant), the original plan was to apply a co-design approach with a service user committee that would contribute to the design of the events and take on different practical tasks. However, before the design phase was started the staff chose not to involve the service users. They argued that at this stage of the initiative the service users would not have the abilities to deliver any useful input (Focus group in case 1, 2018). Instead, they argued that the service users’ perspectives could be included later, as the programme was not fixed and service users could for example choose between playing games or listening to music (Focus group in case 1, 2018). Implementation proceeded on this basis with the staff responsible for choosing and setting up activities, which the service users could choose between. The consequences were that the Friday events (the co-production initiative) became closely related to the pre-existing Monday and Wednesday events, where the service users could choose to dance or sing together (Observations at Case 1, 2017, 2018, 2019).

In case 2 (emotionally close) co-design was more evident. The purpose of the co-production initiative in case 2 was to increase the service users’ active participation and contribution of resources to different tasks in regards to the daily dinner cooking in the shared kitchen. The aim was to use a co-design approach to increase the service users’ independence and for them to rely less on professional support in this part of their lives. In contrast to case 1 the frontline staff chose to interview the service users about what was important to them in regard to changing the cooking routine at the housing complex (Focus group in case 2, 2018). Originally the staff had discussed the possibility of having the food or parts of it delivered, to ease the cooking tasks. In the interviews the service users said that they enjoyed the cooking time and the homemade meals, as a result of which the idea of food delivery was discarded. Instead a peer-to-peer initiative for the service users was created to minimize the work tasks and enable them to have a home-cooked meal (Focus group in case 2, 2018).

However, the implementation in case 2 did not go as planned. After trying to implement the new cooking routine, the frontline staff said that they themselves still did most of the cooking and that the service users would not ‘take on the responsibilities’, despite the frontline staff attempting to individualize, break up and simplify the service users’ tasks (Focus group in case 2, 2019). The staff reflected on how they felt responsible in different matters regarding the cooking processes e.g. that the food was served on time and that the service users would choose the food instead of unhealthy take-aways. These feelings of responsibility sometimes prevented the frontline staff from involving the service users. Instead the frontline staff would increase their own input to the service delivery. An employee describes it this way: ‘ … It can be a stress factor for us as employees and that is why we come to take over [the cooking], because we need to get the food served’ (Female employee in case 2, Focus group 2018).

Comparing the two cases can give us insight into the role of emotional display in creating a conducive environment for co-production. In particular, we link different types of emotional display to different stages of co-production failure – one at the design phase and one at the delivery phase. In case 1, the frontline staff failed to undertake co-design or co-delivery. They did not reflect on if or how they could involve service users, e.g. by simplifying co-production tasks (Steen and Tuurnas 2018). Nor did it appear that they differentiated between the skills of users or viewed them as individuals with different strengths and resources. In case 2, the frontline workers’ more authentic and naturalistic relationships with the service users appeared to motivate them to undertake a participatory co-design of the new cooking arrangements. However, the emotional bond between staff and service users in case 2 may have become a barrier in the delivery phase, with frontline staff being unable to step back from their responsibility for delivery. This could explain the implementation difficulties and lack of co-delivery found in case 2. This points towards the finding that too much emotional closeness may become a barrier for co-delivery due to encouragement of dependence and over-attachment.

Discussion and conclusion – what is the relationship between emotional display rules and co-production?

Here we have investigated co-production in two settings with different emotional display rules, contributing insights about emotion management to debates about the success of co-production. We found that co-design was more evident in the case where staff displayed close emotional links with users than in the case where distance was the dominant norm among the staff. In our data, the site where staff were most keen to emphasize barriers and distance was also the site where they did not trust residents to become involved in the co-design process. In the site where staff display rules were more friendly and naturalistic, there was evidence of co-design. However, this did not lead to effective co-delivery of the service. In both cases, staff perceptions of the risks of failure led to them discontinuing the co-production efforts.

We can’t make causal claims from the data, but by studying the two aspects together, we can make a contribution to theorizing the relationship between emotional display and co-production. Based on this study’s finding, we propose that if staff establish norms of emotional closeness with service users this has a humanizing effect which dismantles barriers and makes it easier to engage in dialogue about new types of service provision, such as the cooking initiative in case 2. In this explanation it is likely that services which establish display rules linked to emotional warmth and closeness will be more successful at co-design than services based on norms of emotional distance. In case 1, where formal display rules were the norm, staff were not willing to begin the co-design dialogue.

However, at the point of co-delivery of outcomes in case 2, the staff did not manage to include service users. Concerns about wellbeing and service standards came to prevail. We suggest that this is because emotional closeness encouraged a form of protectionism from staff. This was closer to the ‘gift’ model of emotional engagement, described by Bolton (2005) as a ‘philanthropic’ version of emotional labour. Mik-Meyer (2016) demonstrates in her disability research that caregiving can cause dependency and people with disabilities can be trapped in unfortunate roles e.g. helpless persons or children. Frontline staff are known to undertake a range of coping behaviours of which the most common is what Tummers et al. call ‘moving towards clients’, which entails that frontline staff ‘pragmatically adjust to the client’s needs, with the ultimate aim to help them’ (2015, 1108). This could be what we found in case 2, where such coping behaviour becomes a barrier for co-production, contributing to relationships of dependence.

Often co-production processes (design, delivery, and evaluation) are reported all together without separating out the different stages; alternatively, studies solely explore one of the ‘co’ phases of co-production (Van Eijk and Steen 2014; Bovaird et al. 2021; Gheduzzi et al. 2021; Krogstrup and Mortensen 2021). We contribute here an understanding of the importance of separating out barriers to the different phases of co-production, recognizing that the forms of emotional display which foster trust and dialogue in service design may encourage paternalism and risk aversion in service delivery. This calls for studies exploring if the mechanism triggering positive results in one phase might become a barrier in another one, as found in this study.

A second contribution of the paper is to separate out the factors which enhance worker’s emotional closeness to service users from those which enhance co-production. We might expect that duration of the relationship is positively linked to both emotional closeness and effective co-production (Pestoff 2019, 24). Indeed, in case 2 where co-production made most progress the duration of the relationship was longer than case 1. However our findings indicate that duration was a factor in emotional closeness and effective co-design but not in successful co-delivery. This suggests a further avenue of research for those considering the variables underpinning effective co-production initiatives.

A third contribution is more practical, bringing attention towards staff training. Guidelines and communication with service users about co-production needs to be sensitive to the existing emotional context – both the formal rules and the prevalent norms. When organizations overlay existing emotional relationships with injunctions to staff to act more co-productively, as they did in our cases, ambiguities in the existing display rules may be brought to the surface. Staff cannot simply be left alone to navigate these ambiguities as they were in the two settings we studied.

There are limitations to this study which impact on the conclusions that can be drawn. First, the data from the case study was collected over a relatively short time period (summer 2017 to spring 2019) and the case design was not designed as a progress study (Yin 2014; Mortensen 2020). Emotional closeness was observed at the same time as the co-production initiatives were being introduced making it hard to separate out the two variables. More empirical evidence is needed to explore how staff balance closeness and distance, and whether there is a point on the continuum that is ‘just right’ – before emotional closeness tips over into protectiveness. Relational studies of public services have often indicated that more emotionally authentic interactions are a feature of effective delivery (e.g. Guy, Newman, and Mastracci 2014; Jensen and Pedersen 2017; Cottam 2018), but Molines et al. (2020) and Boschma (2005) draw attention to the ‘too much of a good thing’ thesis. In their study of transformational leadership (linked to effective emotional support for others), Molines et al. (2020) identified a tipping point after which such leadership styles led to negative consequences such as burnout. Boschma (2005) divided proximity into five dimensions and explored how too much and too little proximity may be harmful to the desired processes of interactive learning and innovation for each dimension. A topic to explore further is a tipping-point for co-production and whether frontline staff can pursue the right amount of emotional closeness for effective co-production.

Another crucial aspect to explore, which is often neglected in the emotion management literature, is the perspective of service users on the appropriate forms of emotional engagement and display rules. Humphrey writes about the importance of recognizing power imbalances within emotional labour (2013, 90). The literature on emotion management usually implies that display rules are set and policed by staff and subject to pressure from service users – in other words that users want more closeness from frontline staff than they are willing/allowed to give. This brings attention to the dynamics of power in co-production. Kirkegaard and Andersen draw on Goffman and their own work with people in mental health residential facilities to argue that, in relation to co-production, boundary blurring can be ‘a game of pretend’ by the professionals (2018, 829). In case 1, we see this in the ways in which staff mould the co-production initiative so that it mirrors existing service provision. In case 2, the game of pretend could apply to the way in which the preparation of meals quickly reverted to being staff-led despite the co-design process.

Lastly, it is important to recognize that the staff-user dyad – which we have focused on – is only one manifestation of the display rules within an organization. Further work could usefully examine the ways in which other actors – e.g. service commissioners, managers and families – encourage or discourage certain types of intimacy and reliance. Existing literature on this has highlighted that managers, for example, can be nervous about fostering emotional closeness or can encourage it as part of the ‘added value’ of the service (Johnson 2015; Breslin and Wood 2016). Families can welcome the more authentic relationships with staff or can be worried that their family member will get too attached to a transient staff member (Needham, Allen, and Hall 2016). More attention is required on the perspectives of these multiple stakeholders in the context of co-design and co-delivery.

In summary, the contribution of this article is to bring an understanding of emotional display rules into the study of co-production, and to argue for its theoretical relevance and empirical significance. Through sharing data from the Danish sites we can see that different emotional norms operate in different services, and that these need to be considered when exploring how co-production initiatives progress. Our empirical study has explored and described a novel phenomenon, which can help scholars to develop more comprehensive theories in the future. Co-production is always at risk of being a ‘game of pretend’ (Kirkegaard and Andersen 2018) given the pre-existing power dynamics in public services. A focus on relational norms helps us to better understand the scope for co-production to become more than make believe.

Acknowledgments

The authors would like to thank the participants at the IIAS study group on Coproduction of Public Services 2019 and the ‘Learning from the Front Line’ panel at IRSPM 2021 conference for valuable feedback on earlier drafts of this manuscript. In addition, they also would like to thank colleagues at the Institute of Local Government Studies (University of Birmingham) and the anonymous reviewers for their helpful comments.

Disclosure statement

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

Additional information

Funding

This work was supported by the Innovation Fund Denmark [6171-00046B].

Notes on contributors

Nanna Møller Mortensen

Nanna Møller Mortensen is an assistant professor of co-production and evaluation at the research group ‘Capacity Building and Evaluation’ at the Department of Culture and Learning at Aalborg University, Denmark. Her research includes topics such as user participation in evaluation, professionalism, and implementation of co-production in public services. She specialises in co-production and the new roles that are ascribed to both professional and non-professional co-producers.

Catherine Needham

Catherine Needham is Professor of Public Policy and Public Management at the Health Services Management Centre, University of Birmingham, UK. Catherine’s research focuses on public service reform, with a particular focus on social care (including care markets, care systems, personalisation, personal budgets, older people, care workforce). She has also developed work on the public service workforce (roles, skills, capabilities of professionals working in public services). She is Associate Director of the National Institute for Health Research’s School for Social Care Research. She is a Co-Investigator on an Economic and Social Research Council Large Grant on Sustainable Care and leads on a work package on care systems in the four nations of the UK. She has published over 70 peer-reviewed publications and written widely for policy and practice audiences. Her most recent book is Reimagining the Future Public Service Workforce (Springer, 2019).